The Laboratory-Based Hyspex Features of Chlorite As the Exploration Tool for High-Grade Iron Ore in Anshan-Benxi Area, Liaoning Province, Northeast China

applied sciences

Article The Laboratory-Based HySpex Features of Chlorite as the Exploration Tool for High-Grade Iron Ore in Anshan-Benxi Area, Liaoning Province, Northeast China

Dahai Hao 1, Yuzeng Yao 1,*, Jianfei Fu 1, Joseph R. Michalski 2 and Kun Song 1

1 School of Resources and Civil Engineering, Northeastern University, Shenyang 110819, China; [email protected] (D.H.); [email protected] (J.F.); [email protected] (K.S.) 2 Division of Earth and Planetary Science & Laboratory for Space Research, the University of Hong Kong, Hong Kong 999077, China; [email protected] * Correspondence: [email protected]

Received: 29 September 2020; Accepted: 21 October 2020; Published: 23 October 2020

Abstract: Anshan-Benxi area in Liaoning province is an important banded iron formations (BIFs) ore-mining district in China. Chlorite is widely distributed in this area, which is related to BIFs and high-grade iron ore, respectively. A fast and convenient method to identify the type and spatial distribution of different chlorites is crucial to the evaluation of high-grade iron ore in this area. Qidashan iron mine is a typical BIFs deposit, and its BIFs-related high-grade iron ore reserves are the second largest in the area. In this paper, the laboratory-based HySpex-320m hyperspectral imaging was used to study the wall rock in Qidashan iron mine. A hyperspectral imaging processing model was established for mineral identification, mineralogy mapping, and chlorite spectral features extraction. The results show that the wavelength positions of OH, Fe-OH, and Mg-OH absorptions of chlorite in the altered wall rock of high-grade iron ore are between 1400 and 1410, 2260 and 2265, and 2360 and 2370 nm, respectively, which are longer than those around BIFs. The relationship between cations in the octahedral layer of chlorite and the wavelengths of OH, Fe-OH, and Mg-OH indicates that Mg and Mg/(Mg + Fe) are inversely related to these wavelengths, whereas Fe is positively related. The wavelengths appear to be weakly influenced by AlVI. Since the bandpass of hyperspectral imaging systems is usually less than 10 nm, these chlorite wavelength differences can be used as a favorable tool for the high-grade iron ore exploration and the iron resources evaluation in the Anshan-Benxi area.

Keywords: laboratory-based hyperspectral imaging; HySpex; spectral features extraction; chlorite; Qidashan Iron Deposit

1. Introduction Nowadays, hyperspectral remote sensing is a hot topic in the remote sensing field. It was originally proposed for minerals. In the 1980s, based on the study of mineral reflectance spectroscopy models by Clark and Hapke [1–3], the unique advantages of hyperspectral remote sensing in geological applications were gradually recognized by many scholars. Traditional hyperspectral imaging is mostly used in aircraft or satellite platforms and suitable for large-scale geology mapping. On the other hand, emerging terrestrial imaging spectroscopy methods, such as laboratory-based and field-based methods, have been used for various geoscience applications, including mineral exploration [4–7], sedimentology analysis [8,9], mineral identification, and alteration zonation mapping, etc. [10–12]. Compared to those of airborne and satellite-borne, the laboratory-based hyperspectral imaging can

Appl. Sci. 2020, 10, 7444; doi:10.3390/app10217444 www.mdpi.com/journal/applsci Appl. Sci. 2020, 10, 7444 2 of 24 obtain mineralogical maps at centimeter to millimeter scale, thus providing more detailed information about the internal spatial distribution of rock specimens. Anshan-Benxi area, Liaoning province, is one of the most important BIFs ore resource bases in China, among which the Qidashan is interpreted as the second largest one. The iron ores in this area are mainly composed of BIFs and BIFs-related high-grade iron ores. The high-grade iron ore has the characteristics of high iron grade and more industrial economics. Chloritization shows close relations to the high-grade iron ore, which can be regarded as the crucial prospecting criteria [13]. Unfortunately, the chlorite does not merely distribute around the high-grade ores, actually, it is widespread everywhere in the mining district, which is thought to be related to the low-medium grade regional metamorphism. Furthermore, high-grade iron ores are also distributed in other deposits of the Anshan-Benxi area. Examples include the Gongchangling, Yanqianshan, Waitoushan, Dawanggou, Banpanling, and Nanfen iron deposits. Their wall rock alteration consists of chloritization, sericitization, etc., which is similar to the Qidashan [14]. Although many researches such as geophysical exploration, geochemistry, and geochronology have been carried out on these mines, the hyperspectral imaging method, which has the advantages of quick-evaluation of the high-grade iron ore resources, has rarely been conducted. In this paper, the laboratory-based HySpex-320m spectrometer was adopted to study the wall rock of Qidashan Iron Deposit. The alteration minerals, assemblage, texture, and the spectral features of chlorite were identiﬁed. A model of hyperspectral image processing and spectral feature extraction was promoted, and the spectral diﬀerences of two kinds of chlorites extracted from wall rock of BIFs and high-grade iron ore were analyzed. Finally, the relationship between cations of the chlorite octahedral layer and the spectral absorption features was discussed.

2. The Application Principle of Hyperspectral Imaging in Geoscience

2.1. The Causes of Absorption Features of Minerals The bandpass of hyperspectral remote sensing is usually below 10 nm, so the continuous spectra and the spectral details can be obtained. This lays the foundation for the identification and analysis of minerals by hyperspectral remote sensing at the spectral level. It is also the main feature and biggest advantage that distinguishes hyperspectral from multispectral remote sensing. In the spectral range of 350~2500 nm, the spectral absorption features of minerals are the diagnostic of crystal structure and chemical composition. The spectra of different minerals are generally different, even the minerals of the same family also have subtle spectral differences due to chemical composition variations. There are two general processes that cause spectral absorption features of minerals: electronic and vibrational [15–19]. In the visible and near-infrared spectral range (VNIR: 350~1200 nm), the causes of spectral absorption features are electronic processes, including crystal field effects, charge transfer absorptions, conduction bands, and color centers, among which the crystal field effects of Fe2+ and Fe3+ and the charge transfer absorptions of Fe2+-Fe3+ are most common. In the short-wave infrared spectroscopy (SWIR: 1200~2500 nm), the causes of the spectral absorption features are predominated by the vibration process, which is mainly caused by the first overtone of the hydroxyl group (OH) and the combination vibrations involving OH stretching and cation-O-H bending in the octahedron. Common vibrational absorptions include aluminum hydroxyl (Al-OH), 2 magnesium hydroxyl (Mg-OH), water (H2O), and carbonate (CO3 −), etc.

2.2. The Spectral Absorption Features of Chlorite The crystal structure of the chlorite has a 2:1 ratio of tetrahedral (T) to octahedral (O) layers 2+ (talc-like) with edge-sharing M (OH)6 octahedra (brucite-like layer) in the interlayer between each set of 2:1 sheets [20] (Figure1). Chlorite has three octahedral cation sites in octahedral layers, one slightly larger M1 site and two slightly smaller M2 sites [21]. Cation substitution of chlorite is common in M sites, and mineral chemistry can vary signiﬁcantly from endmember composition. The substitution Appl. Sci. 2020, 10, 7444 3 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 3 of 24 of Si by Al occurs in the tetrahedral sheet of of 2:1 layer, layer, whereas the mutual substitution of Mg and Fe usually occurs in the brucite brucite-like-like layer and the 2:1 layer of octahedral sheet [[22].22].

Figure 1 1.. TheThe structure of chlorite.

The spectral absorption features features of of chlorite in in the VNIR are mainly caused by the crystal field field effecteffect of iron iron,, which is the most common transition element in minerals. When the Fe is in octahedral octahedral coordination with with six six identic identicalal ligands, ligands, electrons electrons in in all all five five unfilled unfilled 3d 3d orbitals orbitals are are split split into into eg eandg and t2g energyt2g energy levels levels [23] [ 23(Fig] (Figureure 2).2 For). For example, example, the the broad broad absorption absorption band band of of1.0 1.0–1.1–1.1 μmµ min inchlorite chlorite is 2+ 2+ causedis caused by by the the Fe Fe spinspin-allowed-allowed transition transition between between eg e gandand t2g t2g [18].[18]. This This absorption absorption band band is is also commonly foundfound inin phyllosilicate phyllosilicate minerals minerals such such as sericite.as sericite. Other Other spectral spectral features features of chlorite of chlorite are listed are listedin Figure in Figure3. 3.

Figure 2. (a) Six-coordinated octahedron around Fe and (b) crystal field splitting of Fe2+ 3d orbitals in Figure 2. ((aa)) S Six-coordinatedix-coordinated octahedron around Fe andand ((bb)) crystalcrystal ﬁeldfield splittingsplitting ofof FeFe2+ 3d orbitals in the regular octahedron. The octahedral distortion will cause further splitting. the regular octahedron. The The octahedral octahedral distortion will cause further splitting. Appl. Sci. 2020, 10, 7444 4 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 4 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 4 of 24

Figure 3. The spectral absorption features of chlorite [15,17,18]. Figure 3. TheThe spectral spectral absorption absorption features features of chlorite [15,17,18]. [15,17,18]. The spectral spectral absorption absorption features features of chlorite of chlorite in the in SWIR the SWIRare assigned are assigned to combinations to combinations or overtones or The spectral absorption features of chlorite in the SWIR are assigned to combinations or overtonesof the OH of bound the OH to octahedral bound to octahedral cations (Figure cations4). Chlorite(Figure 4). is characterizedChlorite is characterized by the Fe-OH by absorptionthe Fe-OH overtones of the OH bound to octahedral cations (Figure 4). Chlorite is characterized by the Fe-OH absorptionband near 2250 band nm near (combinations 2250 nm (combinations of O-H stretching of andO-H Fe-OH stretching bending and vibration), Fe-OH bending whose wavelength vibration), absorption band near 2250 nm (combinations of O-H stretching and Fe-OH bending vibration), whoseposition wavelength varies as a functionposition of varies the relative as a function amount of of cationsthe relative in the amount octahedral of cations sheet [24 in]. the Other octahedral spectral whose wavelength position varies as a function of the relative amount of cations in the octahedral sheetfeatures [24]. of Other chlorite spectral are listed features in Figure of chlorite3. are listed in Figure 3. sheet [24]. Other spectral features of chlorite are listed in Figure 3.

Figure 4. Schematic diagram of OH vibrations in octahedral sheet of chlorite. The OH-bending motion Figure 4. Schematic diagram of OH vibrations in octahedral sheet of chlorite. The OH-bending motion isFigure shown 4. inSchematic blue and diagram the OH- ofstretching OH vibrations motion in in octahedral yellow (modified sheet of chlorite.from Bishop The OH-bending[20]). motion is shown in blue and the OH OH-stretching-stretching motion in yellow (modiﬁed(modified from BishopBishop [[20]).20]). 3. Geological Settings and Sampling 3. Geological Geological Settings Settings and Sampling Sampling The Qidashan Iron Deposit, with proven BIFs ore reserves of 1259 Mt and high-grade iron ore The Qidashan Iron Deposit, with proven BIFs BIFs ore ore reserves of 1259 Mt and high high-grade-grade iron ore reserves of 11.45 Mt, lies in the Anshan-Benxi area, Northeastern part of the North China Craton [14]. reserves of 11.45 M Mt,t, lies lies in in the the Anshan Anshan-Benxi-Benxi area, area, Northeastern Northeastern part part of of the the North North China China Craton Craton [14]. [14]. The deposit belongs to metamorphic volcanic-sedimentary iron deposit, and the stratum exposed in The deposit belongs to metamorphic volcanic volcanic-sedimentary-sedimentary iron deposit, and the stratum exposed in the mining area is mainly Archean Anshan group and Proterozoic Liaohe Group (Figure 5a). The the miningmining areaarea is is mainly mainly Archean Archean Anshan Anshan group group and and Proterozoic Proterozoic Liaohe Liaohe Group Group (Figure (Figure5a). The 5a). main The main lithology of Yingtaoyuan formation where the BIFs ore was located is chlorite quartz schist, mainlithology lithology of Yingtaoyuan of Yingtaoyuan formation formation where where the BIFs the ore BIFs was ore located was located is chlorite is chlorite quartz schist,quartz chlorite schist, chlorite talc schist, sericite quartzite, metagranite, amphibolite, and banded iron quartzite. Large chloritetalc schist, talc sericite schist, quartzite, sericite quartzite, metagranite, metagranite, amphibolite, amphibolite and banded, and iron banded quartzite. iron quartzite. Large areas Large of areas of Neoarchean granite are distributed on both the east and west sides of the deposit (Figure 5b). areasNeoarchean of Neoarchean granite aregranite distributed are distrib onuted both on the both east the and east west and sides west of sides the deposit of the deposit (Figure 5(Figureb). 5b). The Qidashan Iron Deposit is a typical BIFs deposit. The BIFs are chemical sedimentary rock The Qidashan Iron Deposit is a typical BIFs deposit. The The BIFs BIFs are are chemical sedimentary rock rock characterized by the commonly alternating presence of iron-rich layers and silica-rich layers [25]. characterized by by the the commonly commonly alternating alternating presence presence of of iron iron-rich-rich layers layers and and silica silica-rich-rich layers layers [25]. [25]. BIFs-related iron deposits in China are mainly composed of magnetite (average 30.35% total Fe) with BIFsBIFs-related-related iron deposits in China are mainly composed of magnetite (average 30.35% total Fe) with subordinate amount of high-grade magnetite ores (total Fe ≥ 50%) [26]. Low-grade iron ores (BIFs) in subordinate amount of high-gradehigh-grade magnetite oresores (total(total FeFe ≥ 50%) [26]. [26]. Low Low-grade-grade iron iron ores (BIFs) in the Qidashan Iron Deposit are composed of quartz band and≥ magnetite band with a width of about the Qidashan I Ironron D Depositeposit are composed of quartz band and magnetite band with a width of about 1 mm. The high-grade iron ores in the BIFs are mainly massive and composed of magnetite, quartz, 1 mm. The high high-grade-grade iron iron ores ores in the BIFs are mainly massive and and composed of of magnetite, magnetite, quartz quartz,, and chlorite. They are considered to have resulted from hydrothermal enrichment of BIFs near the and chlorite. They are considered to have resulted from hydrothermal enrichment of BIFs near the faults, which are consistent with the occurrence of the BIFs layer [14]. The main types of wall rock faults, which are consistent with the occurrence of the BIFs layer [14]. The main types of wall rock alteration of Qidashan Iron Deposit include chloritization, sericitization, biotitization, and pyritization. alteration of Qidashan Iron Deposit include chloritization, sericitization, biotitization, and pyritization. Appl. Sci. 2020, 10, 7444 5 of 24 and chlorite. They are considered to have resulted from hydrothermal enrichment of BIFs near the faults, which are consistent with the occurrence of the BIFs layer [14]. The main types of wall rock alterationAppl. Sci. 2020 of, Qidashan10, x FOR PEER Iron REVIEW Deposit include chloritization, sericitization, biotitization, and pyritization.5 of 24 In this work, 15 samples of BIFs, high-grade iron ore, and wall rocks were collected in the In this work, 15 samples of BIFs, high-grade iron ore, and wall rocks were collected in the mining mining area (Figure5c). All the samples were measured using ASD FieldSpec-3 pro spectrometer. area (Figure 5c). All the samples were measured using ASD FieldSpec-3 pro spectrometer. Among Among which, six representative samples (from sampling location O1 and O3 ) were selected for which, six representative samples (from sampling location ① and ③) were selected for laboratory- laboratory-based hyperspectral imaging. These samples include three for high-grade iron ore (Q-2, based hyperspectral imaging. These samples include three for high-grade iron ore (Q-2, Q-9, and Q- Q-9, and Q-11) and the remnant for BIFs (Q-4, Q-12, and Q-13). 11) and the remnant for BIFs (Q-4, Q-12, and Q-13).

Figure 5. (a) Simplified regional geological map of Anshan area, (b) geological map of Qidashan Iron Figure 5. (a) Simpliﬁed regional geological map of Anshan area, (b) geological map of Qidashan Iron Deposit, and (c) sketch diagram of sampling. Deposit, and (c) sketch diagram of sampling.

4.4. Methods AllAll the the 15 15 rock rock samples samples were were measured measured by the by ASD the ASD FieldSpec-3 FieldSpec pro- spectrometer,3 pro spectrometer, and the and spectra the werespectra carefully were carefully investigated investigated for mineral for identiﬁcationmineral identification (see below). (see Sixbelow). representative Six representative samples usedsamples for hyperspectralused for hyperspectral imaging imaging were then were cut then into rockcut in chipsto rock of chips ca. 6–7 of cmca. (Unpolished)6–7 cm (Unpolished) to facilitate to facilitate image acquisitionimage acquisition and processing. and processing. A laboratory-based A laboratory-based hyperspectral hyperspectral imaging imaging data processingdata processing model model was establishedwas established in this in paperthis paper (Figure (Figure6). All 6). the All operations the operations were completedwere completed in ENVI in ENVI 4.7 and 4.7 IDL and 7.1. IDL 7.1. Appl. Sci. 2020, 10, 7444 6 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 24

Figure 6. The laboratory-based hyperspectral imaging processing model. FigureFigure 6.6. TheThe laboratory-basedlaboratory-based hyperspectralhyperspectral imagingimaging processingprocessing model.model. 4.1.4.1. Data Acquisition 4.1. Data Acquisition TheThe hyperspectralhyperspectral imagingimaging spectrometerspectrometer usedused inin thisthis paperpaper isis HySpex-320mHySpex-320m imagingimaging systemsystem The hyperspectral imaging spectrometer used in this paper is HySpex-320m imaging system (Figure(Figure7 7a).a). TheThe rockrock chipschips werewere runrun sequentiallysequentially throughthrough thethe cameracamera atat aa distancedistance ofof about about 0.5 0.5 m m (Figure 7a). The rock chips were run sequentially through the camera at a distance of about 0.5 m underunder artiﬁcialartificial illumination. illumination. Two Two sensors sensors were were used used in thisin this camera, camera, one ofone which of which operated operated in the VNIRin the under artificial illumination. Two sensors were used in this camera, one of which operated in the andVNIR the and other the in other the SWIR in the regions. SWIR The regions. VNIR The sensor VNIR is conﬁgured sensor is configured to record 125 to bands, record consisting 125 bands, a VNIR and the other in the SWIR regions. The VNIR sensor is configured to record 125 bands, fullconsisting spectrum a full with spectrum a spectral with sampling a spectral interval sampling of 3.7 interval nm between of 3.7 0.4 nm and between 1.0 µm. 0.4 The and spectral 1.0 μm. range The consisting a full spectrum with a spectral sampling interval of 3.7 nm between 0.4 and 1.0 μm. The ofspectral the SWIR range sensor of the is 1.0–2.5SWIR sensorµm with is 1.0 256–2.5 bands, μm andwith the256 spectral bands, resolutionand the spectral is 6.25 resolution nm. Because is 6.25 the spectral range of the SWIR sensor is 1.0–2.5 μm with 256 bands, and the spectral resolution is 6.25 hyperspectralnm. Because the images hyperspectral were obtained images from were the cutobtained surface from of the the rock cut chipssurface at aof close the rock distance, chips the at spatiala close nm. Because the hyperspectral images were obtained from the cut surface of the rock chips at a close resolutiondistance, the is asspatial high asresolution 0.3 mm is 0.3as high mm. as 0.3 mm × 0.3 mm. distance, the spatial resolution ×is as high as 0.3 mm × 0.3 mm.

Figure 7. The instruments used for data acquisition in this paper: (a) HySpex hyperspectral imaging Figure 7. The instruments used for data acquisition in this paper: (a) HySpex hyperspectral imaging Figuresystem7. andThe (b instruments) Analytical usedSpectral for Devices data acquisition (ASD) FieldSpec in this paper:-3 pro (spectrometer.a) HySpex hyperspectral imaging system and (b) Analytical Spectral Devices (ASD) FieldSpec-3 pro spectrometer. system and (b) Analytical Spectral Devices (ASD) FieldSpec-3 pro spectrometer. Point spectroscopy is often used to validate mineralogy mapping results [27]. In this paper, ASD Point spectroscopy is often used to validate mineralogy mapping results [27]. In this paper, ASD FieldSpecPoint-3 spectroscopy pro spectrometer is often was used used to to validateobtain the mineralogy spectra of 15 mapping samples results (Figure [27 7b).]. InThe this sampling paper, FieldSpec-3 pro spectrometer was used to obtain the spectra of 15 samples (Figure 7b). The sampling ASDinterval FieldSpec-3 from 0.35 pro to spectrometer 1.0 μm is 1.4 was nm used and to from obtain 1.0 the to spectra2.5 μm ofis 152.0 samples nm. It (Figurewas fitted7b). with The samplinga contact interval from 0.35 to 1.0 µμm is 1.4 nm and from 1.0 to 2.5 µμm is 2.0 nm. It was fitted with a contact intervalprobe with from a light 0.35 source, to 1.0 whichm is 1.4 ensures nm and stable from illumination 1.0 to 2.5 mconditions is 2.0 nm. during It was spectral fitted withmeasurement. a contact probe with a light source, which ensures stable illumination conditions during spectral measurement. probeThe spectra with a light were source, then dividedwhich ensures by the stable reflectance illumination of the conditions calibration during panel spectral to obtain measurement. absolute The spectra were then divided by the reflectance of the calibration panel to obtain absolute Thereflectance. spectra were At thenleast divided 5 individual by the reflectance measurements of the ofcalibration each sample panel to were obtain taken absolute to reflectance.avoid any reflectance. At least 5 individual measurements of each sample were taken to avoid any Atheterogeneity. least 5 individual measurements of each sample were taken to avoid any heterogeneity. heterogeneity. The samples of Q-2, Q-4, Q-11, and Q-13 were analyzed using the electron probe microanalyzer The samples of Q-2, Q-4, Q-11, and Q-13 were analyzed using the electron probe microanalyzer (EPMA) to determine the specific chemical composition of chlorite. The JEOL JXA-8230 was taken at (EPMA) to determine the specific chemical composition of chlorite. The JEOL JXA-8230 was taken at Appl. Sci. 2020, 10, 7444 7 of 24

The samples of Q-2, Q-4, Q-11, and Q-13 were analyzed using the electron probe microanalyzer (EPMA)Appl. Sci. 20 to20, 10 determine, x FOR PEER the REVIEW specific chemical composition of chlorite. The JEOL JXA-82307 wasof 24 taken at the Key Laboratory of Mineralogy and Metallogeny, Guangzhou Institute of Geochemistry, the Key Laboratory of Mineralogy and Metallogeny, Guangzhou Institute of Geochemistry, Chinese Chinese Academy of Sciences. The operating conditions were acceleration voltage of 15 kV, 20 nA Academy of Sciences. The operating conditions were acceleration voltage of 15 kV, 20 nA beam beam current, and 1 µm beam diameter. The analysis standards include tugtupite, orthoclase, albite, current, and 1 μm beam diameter. The analysis standards include tugtupite, orthoclase, albite, BaF2, BaF , kaersutite, diopside, rhodonite, rutile, magnetite, and olivine. kaersutite,2 diopside, rhodonite, rutile, magnetite, and olivine. 4.2. The HySpex Data Preprocessing 4.2. The HySpex Data Preprocessing During data collection, a white reference panel and a gray calibration panel (both with measured reflectanceDuring spectra) data collection, were placed a white in frontreference of the panel samples and a for gray calibration calibration purposes. panel (both The with white measured panel is reflectancemade of a materialspectra) thatwere reflects placed almostin front 100% of the in samples the spectral for calibration range of 0.35–2.5 purposes.µm. The An white average panel dark is madecurrent of framea material was subtractedthat reflects from almost frames 100% of in each the image,spectral and range the of Flat-Field 0.35–2.5 Methodμm. An calculatesaverage dark the currentrelative frame reflectance was subtracted by dividing from the frames spectral of value each ofimage, each pixeland the by Flat the- valueField Method of the white calculates reference the relativepanel to reflectance convert the by raw dividing data to the reflectance. spectral value of each pixel by the value of the white reference panelThe to convert reflectance the dataraw data should to reflectance. be filtered to improve the S/N. The Savitzky–Golay filtering (SG) is a widelyThe used reflectance technique da inta modernshould be hyperspectral filtered to improve remote sensingthe S/N. studies The Savitzky [28–30].–Golay The principle filtering of(SG) SG is athe widely least squareused technique filtering based in modern on moving hyperspectral window [remote31]. The sensing moving studies window [28 contains–30]. The the principle point to beof SGsmoothed is the least in the square center filtering of the windowbased on with moving multiple window adjacent [31]. The points moving on either window side ofcontains the center the point.point Allto be the smoothed points in in the the window center of are the used window to perform with multiple the least adjacent squares points fitting, on but either only side the centralof the center point (point.i = 0) All is smoothed the points (Figure in the 8window). It works are especiallyused to perform well when the least the typical squares band fitting, of the but signal only the is narrow, central pointand the (i height= 0) is smoothed and width ( ofFigu there spectrum 8). It works can especially well be preserved. well when The the filtering typical ebandffect ofof thethe spectrum signal is narrow,varies with and the the size height of the and moving width of window the spectrum and the can order well of be the preserved. polynomial. The Afiltering larger windoweffect of canthe spectrumproduce better varies smoothing with the e sizeffect, of but the may moving lose the window tiny absorption and the featuresorder of (Figure the polynomial.9b). A higher A larger order windowpolynomial can can produce achieve better a better smoothing smoothing effect, without but may attenuation lose the oftiny data absorption features [features32,33]. In (Figure this paper, 9b). Athe higher moving order window polynomial size of 5 can points achieve and 2 a polynomial better smoothing order is without adopted attenuation for the hyperspectral of data features images (Figure[32,33]. 9Ina). this paper, the moving window size of 5 points and 2 polynomial order is adopted for the hyperspectral images (Figure 9a).

Figure 8. The principle of Savitzky–Golay filtering. Figure 8. The principle of Savitzky–Golay ﬁltering. Appl. Sci. 2020, 10, 7444 8 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 24

Figure 9. The results of Savitzky–Golay filtering: (a) smooth result of 5 points and 2 order and (b) Figure 9. The results of Savitzky–Golay filtering: (a) smooth result of 5 points and 2 order and smoothFigure result9. The of results 6 points of Savitzkyand 2 order.–Golay A largerfiltering: window (a) smooth will minimizeresult of 5the points subtle and spectral 2 order features and (b ) (b) smooth result of 6 points and 2 order. A larger window will minimize the subtle spectral features (bluesmooth box). result of 6 points and 2 order. A larger window will minimize the subtle spectral features (blue(blue box). box). 4.3.4.3. The The HySpex HySpex Data Data Processing Processing 4.3. The HySpex Data Processing TheThe HySpex HySpex data data processing processing mainly mainly includes includes two twoparts: parts: mineralogy mineralogy mapping mapping and chlorite and chlorite spectral The HySpex data processing mainly includes two parts: mineralogy mapping and chlorite absorptionspectral absorption features extraction features extraction (Figure6 b).(Figure 6b). spectral absorption features extraction (Figure 6b). High-gradeHigh-grade iron iron ore ore is is mainly mainly composed composed of magnetite and and pyrite pyrite whose whose spectral spectral absorption absorption featuresHigh are-grade concentrated iron ore in is VNIR. mainly Band composed math is of a magnetitecommon remote and pyrite sensing whose method spectral to distinguish absorption features are concentrated in VNIR. Band math is a common remote sensing method to distinguish thefeatures opaque are metal concentrated minerals in[34]. VNIR. High Band-grade math iron is ore a common is mainly remote comp osedsensing of magnetitemethod to and distinguish pyrite. the opaque metal minerals [34]. High-grade iron ore is mainly composed of magnetite and pyrite. Accordingthe opaque to metalthe spectra minerals of pyrite [34]. andHigh magnetite-grade iron in VNIRore is imagemainly ( Figurecomposed 10a), of the magnetite model is andexpressed pyrite. AccordingAccording to theto the spectra spectra of of pyrite pyrite and and magnetite magnetite in in VNIRVNIR imageimage ( (FigureFigure 10a),10a), the the model model is is expressed expressed as K1 (Band 10–Band 50). The K1 is the slope of the spectral reflectance from Band 10 to Band 50. as Kas1 K(Band1 (Band 10–Band 10–Band 50). 50). The TheK K11 isis thethe slopeslope ofof thethe spectral reflectance reflectance from from Band Band 10 10 to to Band Band 50. 50. Obviously, K1 of pyrite is greater than 0, whereas less than 0 for magnetite. BIFs ore is mainly Obviously, K of1 pyrite is greater than 0, whereas less than 0 for magnetite. BIFs ore is mainly composedObviously, of1 Kmagnetite of pyrite and is quartz, greater and than the 0, slope whereas of the less spectral than reflectance 0 for magnetite. from Band BIFs 20 ore to isBand mainly 70 composedarecomposed quite ofdifferent magnetiteof magnetite (Figure and and 10b). quartz, quartz, The and model and the the distinguishing slope slope of of thethe spectralspectral magnetite reflectance reflectance and quartz from from is Bandthen Band expressed20 20 to to Band Band as 70 70 are quite different (Figure 10b). The model distinguishing magnetite and quartz is then expressed as areK quite2 (Band di 20fferent–Band (Figure 70) < − 0.510b). for Thequartz model and distinguishing−0.5 < K2 < 0 (Band magnetite 20–Band and 70) quartzfor magnetite. is then expressed as K (BandK2 (Band 20–Band 20–Band 70) 70)< < 0.5−0.5 for for quartz quartz andand −0.5 << KK2 < <0 (Band0 (Band 20 20–Band–Band 70) 70) for formagnetite. magnetite. 2 − − 2

Figure 10. The visible and near-infrared spectral range (VNIR) spectra of magnetite, pyrite, and FigurequartzFigure 10.: ( a10.The) spectra The visible visible of andpyrite and near-infrared nearand magnetite-infrared spectral spectral in Q-2 range anranged ( (VNIR)b )( VNIRspectra spectra) spectra of quartz of of magnetite, magnetite,and magnetite pyrite, pyrite in, and Qand-12. quartz: quartz: (a) spectra of pyrite and magnetite in Q-2 and (b) spectra of quartz and magnetite in Q-12. (a) spectra of pyrite and magnetite in Q-2 and (b) spectra of quartz and magnetite in Q-12. The main diagnostic spectral absorption features of altered minerals such as chlorite, muscovite, andThe biotiteThe main main are diagnostic located diagnostic in spectralSWIR. spectral “MNF absorption absorption-PPI-n Dimensional features features ofof alteredaltered,” a common minerals hyperspectral such such as as chlorite, chlorite, remote muscovitemuscovite, sensing , andprocessingand biotite biotite are meth are located locatedod [35 in,36] in SWIR. ,SWIR. was “MNF-PPI-nused “MNF to extract-PPI-n Dimensional,” Dimensionalthe endmember,” aa spectra common from hyperspectral hyperspectral SWIR images. remote remote According sensing sensing processingtoprocessing the previous method meth researchod [35 [35,36],,36] was on, was Qidashan used used to extractto Ironextract Deposit, the the endmember endmember the spectra spectra spectra of minerals from from SWIR SWIR that images. images. may exist According According in the to to the previous research on Qidashan Iron Deposit, the spectra of minerals that may exist in the Appl. Sci. 2020, 10, 7444 9 of 24 theAppl. previous Sci. 2020 research, 10, x FOR onPEER Qidashan REVIEW Iron Deposit, the spectra of minerals that may exist in the mining9 of 24 area were selected from USGS spectral library as the customer-defined spectral library. The Spectral mining area were selected from USGS spectral library as the customer-defined spectral library. The Angle Mapping (SAM), which has proven to be an effective classification method for hyperspectral Spectral Angle Mapping (SAM), which has proven to be an effective classification method for images [10,11,28], was adopted to determine the spectral similarity between the endmember spectra hyperspectral images [10,11,28], was adopted to determine the spectral similarity between the andendmember the customer-defined spectra and spectra the customer by calculating-defined thespect anglera by between calculating the the spectra angle so between that the the endmember spectra spectraso that can the be endmember identified [37 spectra].The operator can be identified then checked [37].The the resultsoperator of thethen SAM checked based the on results geological of the and spectralSAM knowledgebased on geological to ensure and that spectral the endmember knowledge spectra to ensure were that not misidentified.the endmember The spectra maximum were not angle of 0.15misidentified. rad was used The asmaximum the threshold angle of for 0.15 the rad SAM was mineralogy used as the mapping threshold in for this the paper. SAM Bymineralogy combining knowledgemapping ofin mineralthis paper. spectroscopy, By combining pixels knowledge with spectral of mineral angle spectroscopy, higher than thresholdpixels with were spectral determined angle byhigher manual than inspection threshold until were no determined unclassified by pixels manual remained. inspection until no unclassified pixels remained. TheThe result result map map of of mineral mineral classification classification waswas convertedconverted to vector, vector, then then the the chlorite chlorite was was extracted extracted andand other other minerals minerals were were masked masked to to obtain obtain the the purepure chloritechlorite thematic map. map. The The continuum continuum-removal-removal waswas used used to to extract extract the the absorptionabsorption wavelength wavelength position position and and depth depth of chlorite. of chlorite. The continuum The continuum is a is a“shell “shell”” wrapped wrapped around around the the spectrum spectrum (Figure (Figure 1 1).11 ). The The continuum continuum-removal-removal can can remove remove the the “background“background spectrum” spectrum to” highlight to highlight the subtle the subtle absorption absorption features, features, which which is beneficial is beneficial to the extraction to the andextraction comparison and ofcomparison the spectral of the features spectral of difeaturesfferent of minerals. different minerals.

FigureFigure 11. 11.Comparison Comparison ofof originaloriginal and continuum-removalcontinuum-removal spectrum. spectrum.

5. Results5. Results

5.1.5.1. The The Spectra Spectra of of Point Point Measurement Measurement TheThe point point spectra spectra of of the the samples samples used used forfor imagingimaging are shown in in Figure Figure 12. 12 .Figure Figure 55 illustr illustratesates thatthat the the Q-9 Q- was9 was taken taken from from mica-quartz mica-quartz schistschist zone,zone, whereas the the Q Q-4,-4, Q Q-11,-11, and and Q Q-13-13 were were taken taken fromfrom chloritite chloritite and and chlorite-quartz chlorite-quartz schist schist zone. zone. Correspondingly Correspondingly their their spectra spectra manifest manifest obviously obviously similarsimilar features features as as those those of of muscovite muscovite and and chlorite,chlorite, respectively.respectively. The The Q Q-2-2 and and Q Q-12-12 were were selected selected from from high-gradehigh-grade iron iron ore ore and and BIFs, BIFs, which which are are characterized characterized by lower reflectivity reﬂectivity and and their their main main absorption absorption featuresfeatures are are concentrated concentrated in in VNIR VNIR range. range. Chlorite is marked by diagnostic absorptions around 2250 and 2350 nm [6,15,17,18]. By mutual Chlorite is marked by diagnostic absorptions around 2250 and 2350 nm [6,15,17,18]. By mutual verification with hyperspectral imaging, the samples containing chlorite from wall rock of BIFs (Q-4, veriﬁcation with hyperspectral imaging, the samples containing chlorite from wall rock of BIFs (Q-4, Q-8, Q-13, and Q-15) and high-grade iron ore (Q-3, Q-6, Q-11, and Q-14) were selected, and the wavelength positions around 2250 and 2350 nm were extracted (Figure 13). Appl. Sci. 2020, 10, 7444 10 of 24

Q-8, Q-13, and Q-15) and high-grade iron ore (Q-3, Q-6, Q-11, and Q-14) were selected, and the wavelength positions around 2250 and 2350 nm were extracted (Figure 13). Appl.Appl. Sci. Sci. 20 202020, 10, 10, x, x FOR FOR PEER PEER REVIEW REVIEW 10 of 24

FigureFigureFigure 12.12. 12. The The point point spectra spectra of of the the samples: samples: ((aa)) samplessamples nearnear high-gradehighhigh--gradegrade ironiron oreore andandand (((bbb))) samplessamplessamples nearnearnear bandedbandedbanded ironiron iron formationsformations formations (BIFs).( BIFs(BIFs).) .

FigureFigure 13. 13. Scatter Scatter diagram diagram of of the the wavelengthwavelength positionspositions nearnear 22502250 andand 23502350 nmnm eextractedxtracted from the Figure 13. Scatter diagram of the wavelength positions near 2250 and 2350 nm extracted from the pointpoint spectra. spectra. point spectra.

5.2.5.2.5.2. Minerals Minerals IdentiﬁcationIdentification Identification fromfrom from thethe the HySpexHySpex HySpex DataData Data AccordingAccordingAccording toto to thethe the HySpexHySpex HySpex data datadata processing processingprocessing model, model,model, thethethe endmemberendmemberendmember spectralspectralspectral librarylibrarylibrary isis shownshown inin FigureFigureFigure 14 14. 14.. The The The chlorite, chlorite, chlorite, sericite, sericite, sericite, calcite, calcite, calcite, biotite, biotite, biotite, tourmaline, tourmaline, tourmaline, quartz, quartz, quartz, magnetite, magnetite, magnetite, and pyrite andand were pyrite extracted were fromextractedextracted the HySpex from from the the images. HySpex HySpex images. images. Appl. Sci. 2020, 10, 7444 11 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 11 of 24

FigureFigure 14. 14.The The endmemberendmember spectral spectral library library for the for short the-wave short-wave infrared infrared spectroscopy spectroscopy (SWIR) mineral (SWIR) mineralmapping. mapping.

ChloriteChlorite is ais widespread a widespread alteration alteration mineral mineral inin thethe QidashanQidashan mining mining district, district, which which manifests manifests diagnosticdiagnostic absorptions absorptions around around 2250 2250 and and 2350 2350 nm nm [6[6,17].,17]. Spectral absorption absorption features features ranging ranging from from 2245 to 2265 nm commonly result from Fe-OH vibration. Absorptions caused by Mg-OH vibration 2245 to 2265 nm commonly result from Fe-OH vibration. Absorptions caused by Mg-OH vibration commonly occur in the 2330~2370 nm range [16]. The specific wavelength position of Fe-OH commonly occur in the 2330~2370 nm range [16]. The specific wavelength position of Fe-OH absorption absorption feature varies with Fe and/or Mg content in octahedral sheet, i.e., Mg-rich chlorite is featuregenerally varies between with Fe and2245/or and Mg 2255 content nm, whereas in octahedral Fe-rich sheet, chlorite i.e., tends Mg-rich to be chlorite around is2265 generally nm. The between Mg- 2245OH and wavelength 2255 nm,whereas for the pure Fe-rich chlorite chlorite rich in tends Mg is to around be around 2330 nm 2265 and nm. for The Fe-rich Mg-OH chlorite wavelength is 2365 nm for the pure[24]. Some chlorite scholars rich in believe Mg is that around the OH 2330 absorption nm and fornear Fe-rich 1400 nm chlorite is also ispositively 2365 nm correlated [24]. Some with scholars the believetotal that Fe content the OH in absorption chlorite [17]. near 1400 nm is also positively correlated with the total Fe content in chlorite [White17]. mica is characterized by a major Al–OH band near 2200 nm and a subordinate absorption bandWhite near mica 2340 is characterizednm. The observed by awavelength major Al–OH of the band Al– nearOH band 2200 covers nm and a wide a subordinate range from absorption 2185 to band2225 near nm. 2340 This nm. wavelength The observed range wavelength indicates a compositional of the Al–OH variation band covers from aparagonitic wide range (Al from-rich2185 and to 2225Na nm.-bearing, This wavelength with Al–OH rangeband near indicates 2190 nm) a compositional through muscovite variation (K-rich from and paragoniticAl-rich, Al–OH (Al-rich band at and Na-bearing,around 2200 with nm) Al–OH to phengitic band near (Al-poor 2190 and nm) Mg through-rich or muscoviteFe2+-rich, with (K-rich Al–OH and band Al-rich, > 2210 Al–OH nm) [24,38]. band at aroundThe 2200wavelength nm) to phengiticof the Al-OH (Al-poor absorption and Mg-richin the Qidashan or Fe2+ -rich,HySpex with data Al–OH range bandfrom 2196> 2210 to nm)2205 [nm24,,38 ]. Thesuggestive wavelength of ofmuscovite the Al-OH affinity. absorption in the Qidashan HySpex data range from 2196 to 2205 nm, Carbonate minerals mainly include calcite, magnesite, and dolomite. They can be distinguished by suggestive of muscovite affinity. the absorption feature around 2300 nm. The vibration absorption band of CO32- in carbonate is located Carbonate minerals mainly include calcite, magnesite, and dolomite. They can be distinguished around 2340 nm for calcite, 2300 nm for magnesite, and 2320 nm for dolomite [392-–41]. The main by theabsorption absorption band feature of biotite around is the vibration 2300 nm. of Mg The-OH vibration located around absorption 2340 nm band [16]. of Tourmaline CO3 in carbonate produces is locateddiagnostic around absorptions 2340 nm for at calcite, 2200 and 2300 2360 nm nm, for which magnesite, are mainly and 2320 caused nm by for the dolomite Fe-OH [and39– 41Mg].-OH The [6]. main absorptionThe spectrum band of of biotite quartz is is the flat vibration and no diagnostic of Mg-OH absorption located around feature 2340 in SWIR nm [except16]. Tourmaline for OH and produces H2O; diagnosticMagnetite absorptions and pyrite atare 2200 opaque and minerals 2360 nm, that which are known are mainly for low causedreflectivity by theand Fe-OHa few spectral and Mg-OH features [6]. Thein spectrum the SWIR. of Therefore, quartz is the flat VNIR and no data diagnostic were used absorption for mapping feature of these in opaque SWIR exceptminerals. for OH and H2O; MagnetiteFigure and 15 pyrite shows are the opaque results of minerals mineral thatmapping are knownfrom the for HySpex low reflectivitydata. Samples and near a fewBIFs spectralwere featuresQ-4, inQ-12 the, and SWIR. Q-13 Therefore, and near high the- VNIRgrade iron data ore were were used Q-2, for Q- mapping9, and Q-11. of Q these-9 is sericite opaque quartz minerals. schist whoseFigure main 15 shows minerals the are results quartz of and mineral sericite, mapping accompanied from the by HySpextourmaline data. veinlets; Samples Q-12 near and BIFsQ-2 are were Q-4,BIFs Q-12, and and high Q-13-grade and iron near ore high-grade, respectively. iron Except ore were for magnetite,Q-2, Q-9, and Q-2 Q-11. is accompanied Q-9 is sericite by strong quartz schistchloritization whose main and minerals pyritization, are quartz whereas and Q sericite,-12 is mainly accompanied composed by of tourmaline banded quartz veinlets; and Q-12 magnetite and Q-2 Appl. Sci. 2020, 10, 7444 12 of 24 are BIFs and high-grade iron ore, respectively. Except for magnetite, Q-2 is accompanied by strong chloritizationAppl. Sci. 2020 and, 10, x pyritization, FOR PEER REVIEW whereas Q-12 is mainly composed of banded quartz and12 magnetite of 24 with a width of about 1 mm. Q-4, Q-11, and Q-13 contain both chlorite and quartz. Q-11 is mainly with a width of about 1 mm. Q-4, Q-11, and Q-13 contain both chlorite and quartz. Q-11 is mainly composed of chlorite with minor quartz and pyrite. Q-4 and Q-13 contain more quartz than chlorite, composed of chlorite with minor quartz and pyrite. Q-4 and Q-13 contain more quartz than chlorite, whichwhich belong belong to chlorite to chlorite quartz quartz schist. schist. Moreover, Moreover, Q Q-4-4 and and Q Q-13-13 also also contain contain small small amounts amounts of calcite of calcite and biotite.and biotite. The The Leica Leica DM4P DM4P microscope microscope waswas used to to check check the the classification classification results results (Figure (Figure 16). 16).

Figure 15. The result of hyperspectral mineral mapping: (a) the photo and mineral map of Q-2; (b) Figure 15. The result of hyperspectral mineral mapping: (a) the photo and mineral map of Q-2; (b) the the photo and mineral map of Q-4; (c) the photo and mineral map of Q-9; (d) the photo and mineral photomap and of mineral Q-11; (e map) the photo of Q-4; and (c) mineral the photo map and of Q mineral-12; (f) the map photo of Q-9; and (minerald) the photo map of and Q-13. mineral map of Q-11; (e) the photo and mineral map of Q-12; (f) the photo and mineral map of Q-13. Appl. Sci. 2020, 10, 7444 13 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 13 of 24

Figure 16. Photomicrographs of the samples (plane-polarized light): (a) the photomicrograph of Q-2; Figure 16. Photomicrographs of the samples (plane-polarized light): (a) the photomicrograph of Q-2; (b) the photomicrograph of Q-4; (c) the photomicrograph of Q-9; (d) the photomicrograph of Q-11; (b) the photomicrograph of Q-4; (c) the photomicrograph of Q-9; (d) the photomicrograph of Q-11; (e) the photomicrograph of Q-12; (f) the photomicrograph of Q-13. (e) the photomicrograph of Q-12; (f) the photomicrograph of Q-13.

5.3.5.3. Mapping Absorption Features ofof ChloriteChlorite fromfrom thethe HySpexHySpex DataData AccordingAccording to the the HySpex HySpex data data processing processing model, model, the the wavelength wavelength position position of OH, of OH,Fe-OH Fe-OH,, and andMg-OH Mg-OH absorptions absorptions of chlorite of chlorite were mapped. were mapped. Figure 17 Figure illustrates 17 illustrates that there that are obviously there are obviouslytwo kinds twoof chlorite, kinds ofwhich chlorite, have whichdifferent have absorption diﬀerent wavelength absorptions of wavelengths OH, Fe-OH of, and OH, Mg Fe-OH,-OH. The and absorption Mg-OH. Thewavelengths absorption of wavelengthsOH, Fe-OH, and of OH, Mg Fe-OH,-OH related and Mg-OHto high-grade related ore to (Q high-grade-2 and Q-11) ore are (Q-2 mostly and Q-11) between are mostly1400 and between 1410, 2260 1400 and and 2265 1410,, and 2260 2360 and and 2265, 2370 and, nm 2360, respectively, and 2370, whereas nm, respectively, those related whereas to BIFs those (Q- related4 and Q to-13) BIFs are (Q-4between and 1390 Q-13) and are 1400, between 2245 and 1390 2255 and, and 1400, 2330 2245 and and 2350 2255, nm. andThis 2330 result and is consistent 2350 nm. Thiswith resultthe result is consistent of ASD spectrometer. with the result of ASD spectrometer. Appl. Sci. 2020, 10, 7444 14 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 14 of 24

Figure 17. Results of the absorption wavelength position mapping of chlorite: (a) mineral map; (b) Figure 17. Results of the absorption wavelength position mapping of chlorite: (a) mineral map; wavelength of OH; (c) wavelength of Fe-OH; (d) wavelength of Mg-OH. (b) wavelength of OH; (c) wavelength of Fe-OH; (d) wavelength of Mg-OH.

InIn order order to to verify verify the the results results of of the thespectral spectral analysis,analysis, the chlorite chlorite in in the the Q Q-2,-2, Q Q-11,-11, Q Q-4,-4, and and Q Q-13-13 was analyzed by EPMA (Table 1). Note that when Na2O > 1.0%, CaO > 0.5%, or K2O > 0.5%, data were was analyzed by EPMA (Table1). Note that when Na 2O > 1.0%, CaO > 0.5%, or K2O > 0.5%, data were discarded as they suggested severe contamination by adjacent minerals. The structural formula of discarded as they suggested severe contamination by adjacent minerals. The structural formula of chlorite was calculated based on 28 oxygen atoms (Table 2). The distribution of the corresponding chlorite was calculated based on 28 oxygen atoms (Table2). The distribution of the corresponding wavelengths and the cations in the octahedral sheet of chlorite are summarized in Figure 18. wavelengths and the cations in the octahedral sheet of chlorite are summarized in Figure 18. Figure 18 illustrates that there are obviously two types of chlorites (Figure 18a): the wavelengths Figure 18 illustrates that there are obviously two types of chlorites (Figure 18a): the wavelengths of OH, Fe-OH, and Mg-OH of chlorite related to high-grade iron ore (Q-2 and Q-11) are located at of OH, Fe-OH, and Mg-OH of chlorite related to high-grade iron ore (Q-2 and Q-11) are located 1405~1410, 2260~2265, and 2360~2370 nm, respectively, whereas the corresponding wavelengths at 1405~1410, 2260~2265, and 2360~2370 nm, respectively, whereas the corresponding wavelengths related to BIFs (Q-4 and Q-13) are all shorter. On the other hand, similar conclusions can be drawn relatedfrom the to BIFs molecular (Q-4 and formula Q-13) of are the all chlorite: shorter. the On Fe thecontent other in hand, Q-2 and similar Q-11 conclusionsis higher than can that be in drawn Q-4 fromand the Q-13, molecular whereas formula the Mg of is the just chlorite: the opposite the Fe (Figure content 18b). in Q-2 The and AlVI Q-11in these is higher two samples than that are in VI Q-4approximately and Q-13, whereas equal. the Mg is just the opposite (Figure 18b). The Al in these two samples are approximatelyThe absorption equal. depth is also one of the spectral features. The absorption depths of OH, Fe-OH, andThe Mg absorption-OH bands depthof chlorite is also in oneQ-11 of and the Q spectral-13 are illustrat features.ed Thein Figure absorption 19. In depthscomparison of OH, with Fe-OH, the andwavelength, Mg-OH bands there ofis no chlorite obvious in Q-11difference and Q-13between are illustratedthe absorption in Figure depths 19 of. these In comparison samples. with the wavelength, there is no obvious di ﬀerence between the absorption depths of these samples. Appl. Sci. 2020, 10, 7444 15 of 24

Table 1. The electron probe microanalyzer (EPMA) analysis result of the chlorite chemical composition.

Comment K2O Na2O SiO2 TiO2 Al2O3 FeO MgO MnO CaO Cl P2O5 Total Q-2-01 0.01 0.01 22.39 0.05 19.16 35.87 6.60 0.19 0 0.01 0.02 84.31 Q-2-02 0.00 0.01 22.65 0.04 19.20 36.63 6.46 0.15 0.04 0.03 0.00 85.21 Q-2-03 0.00 0.00 22.87 0.04 19.15 35.80 7.13 0.17 0.02 0.02 0.00 85.20 Q-2-04 0.01 0.02 23.40 0.07 18.71 36.48 7.13 0.13 0.03 0.01 0.01 86.00 Q-2-05 0.03 0.02 22.76 0.03 19.51 36.35 6.93 0.19 0.01 0.01 0.03 85.87 Q-2-06 0.00 0.03 22.84 0.03 18.56 36.76 6.97 0.17 0.20 0.01 0.01 85.58 Q-2-07 0.00 0.00 22.82 0.06 19.66 37.28 6.88 0.20 0.00 0.01 0.00 86.91 Q-2-08 0.01 0.01 23.21 0.03 18.72 37.51 6.18 0.15 0.00 0.01 0.01 85.84 Q-2-09 0.00 0.00 23.74 0.06 18.78 36.62 7.12 0.13 0.00 0.01 0.00 86.46 Q-2-10 0.00 0.00 23.28 0.05 19.83 37.13 7.00 0.15 0.00 0.00 0.00 87.44 Q-2-11 0.01 0.01 23.73 0.07 18.64 35.53 8.02 0.17 0.00 0.01 0.00 86.19 Q-11-01 0.01 0.03 22.81 0.03 19.77 37.11 7.05 0.15 0.00 0.01 0.00 86.97 Q-11-02 0.00 0.00 22.89 0.05 19.64 36.96 7.17 0.19 0.00 0.01 0.00 89.91 Q-11-03 0.01 0.00 23.18 0.04 19.53 37.56 6.77 0.16 0.00 0.00 0.00 87.25 Q-11-04 0.01 0.01 23.00 0.02 20.11 36.38 7.62 0.15 0.00 0.02 0.01 87.33 Q-11-05 0.01 0.01 22.68 0.06 18.23 37.22 6.33 0.13 0.23 0.00 0.01 84.93 Q-11-06 0.00 0.01 22.94 0.03 19.42 37.14 7.08 0.21 0.00 0.01 0.00 86.84 Q-11-07 0.02 0.04 23.31 0.03 19.54 36.58 7.08 0.16 0.01 0.02 0.00 86.79 Q-11-08 0.00 0.00 22.99 0.02 19.98 36.25 7.35 0.20 0.00 0.02 0.00 86.81 Q-11-09 0.00 0.00 23.39 0.01 19.86 36.47 7.20 0.19 0.00 0.00 0.00 87.12 Q-11-10 0.00 0.00 17.84 0.03 13.74 46.61 6.02 0.18 0.06 0.01 0.01 84.5 Q-11-11 0.00 0.00 22.96 0.05 19.94 36.52 7.05 0.12 0.05 0.01 0.01 86.71 Q-11-12 0.00 0.01 22.61 0.05 19.98 36.43 7.11 0.16 0.00 0.01 0.00 86.36 Q-11-13 0.00 0.00 23.02 0.03 19.49 37.19 6.46 0.16 0.00 0.01 0.00 86.36 Q-11-14 0.00 0.03 22.75 0.05 19.89 37.19 7.00 0.18 0.00 0.03 0.00 87.12 Q-11-15 0.00 0.00 22.75 0.04 20.06 37.17 6.91 0.19 0.00 0.00 0.02 87.14 Q-13-01 0.00 0.00 26.62 0.02 19.87 18.60 19.05 0.40 0.00 0.00 0.00 84.56 Q-13-02 0.03 0.02 26.51 0.04 18.61 18.95 19.12 0.39 0.02 0.00 0.00 83.69 Q-13-03 0.03 0.02 27.01 0.05 19.70 18.84 19.49 0.39 0.00 0.00 0.02 85.55 Q-13-04 0.01 0.00 28.36 0.01 16.89 19.75 20.28 0.38 0.04 0.01 0.01 85.74 Q-13-05 0.03 0.03 27.87 0.04 18.61 17.38 21.20 0.40 0.00 0.01 0.00 85.57 Q-13-06 0.00 0.00 25.65 0.05 20.26 18.84 19.91 0.39 0.03 0.00 0.00 85.13 Q-13-07 0.01 0.03 26.41 0.03 19.55 19.07 19.77 0.38 0.02 0.00 0.00 85.27 Q-13-08 0.01 0.01 27.1 0.02 18.91 18.78 19.87 0.38 0.00 0.01 0.01 85.1 Q-13-09 0.00 0.01 26.29 0.04 20.13 18.99 19.61 0.41 0.00 0.00 0.00 85.48 Q-13-10 0.01 0.01 26.9 0.05 19.7 18.89 19.80 0.36 0.00 0.01 0.01 85.74 Q-13-11 0.00 0.00 27.17 0.04 19.32 18.84 19.31 0.40 0.00 0.01 0.02 85.11 Q-13-12 0.02 0.00 26.82 0.02 19.68 18.36 19.44 0.36 0.00 0.01 0.00 84.71 Q-13-13 0.02 0.01 27.05 0.03 19.50 18.56 19.39 0.42 0.00 0.01 0.00 84.99 Q-13-14 0.08 0.01 27.01 0.03 19.04 18.66 20.55 0.36 0.00 0.00 0.01 85.75 Q-13-15 0.02 0.00 26.67 0.03 19.74 18.92 19.68 0.40 0.00 0.00 0.00 85.46 Q-4-01 0.00 0.06 26.10 0.02 19.49 18.84 19.33 0.39 0.01 0.10 0.01 84.35 Q-4-02 0.06 0.01 30.32 0.11 16.18 18.63 19.50 0.35 0.04 0.01 0.00 85.21 Q-4-03 0.04 0.00 27.34 0.04 19.23 19.02 19.39 0.34 0.00 0.00 0.00 85.46 Q-4-04 0.00 0.02 26.84 0.02 19.27 19.06 19.44 0.39 0.03 0.01 0.00 85.08 Q-4-05 0.00 0.03 28.48 0.03 18.25 17.46 19.67 0.42 0.00 0.00 0.01 84.71 Q-4-06 0.01 0.00 26.74 0.05 19.08 18.98 19.62 0.38 0.00 0.01 0.00 84.87 Q-4-07 0.03 0.10 27.65 0.02 17.96 18.80 19.67 0.37 0.00 0.13 0.00 84.73 Appl. Sci. 2020, 10, 7444 16 of 24

Table 2. Calculation of the chlorite structural formula (based on 28 oxygen atoms).

Comment Si AlIV AlVI Ti Fe3+ Fe2+ Mn Mg Ca Na K Cl Mg/(Fe + Mg) Q-2-01 5.27 2.73 2.60 0.01 0.00 7.11 0.04 2.31 0.00 0.00 0.00 0.01 0.25 Q-2-02 5.28 2.72 2.58 0.01 0.00 7.20 0.03 2.25 0.01 0.01 0.00 0.02 0.24 Q-2-03 5.31 2.69 2.56 0.01 0.00 7.00 0.03 2.47 0.00 0.00 0.00 0.02 0.26 Q-2-04 5.39 2.61 2.48 0.01 0.00 7.07 0.03 2.45 0.01 0.02 0.01 0.01 0.26 Q-2-05 5.25 2.75 2.57 0.01 0.00 7.09 0.04 2.38 0.01 0.02 0.01 0.01 0.25 Q-2-06 5.31 2.69 2.42 0.01 0.00 7.26 0.03 2.42 0.00 0.02 0.00 0.00 0.25 Q-2-07 5.21 2.79 2.53 0.01 0.00 7.22 0.04 2.34 0.00 0.02 0.00 0.01 0.24 Q-2-08 5.40 2.60 2.53 0.01 0.00 7.32 0.03 2.14 0.00 0.00 0.00 0.00 0.23 Q-2-09 5.44 2.56 2.52 0.01 0.00 7.03 0.02 2.43 0.00 0.00 0.00 0.00 0.26 Q-2-10 5.28 2.72 2.58 0.01 0.01 7.09 0.03 2.36 0.00 0.00 0.00 0.01 0.25 Q-2-11 5.42 2.58 2.45 0.01 0.00 6.83 0.03 2.73 0.00 0.00 0.01 0.01 0.29 Q-11-01 5.20 2.80 2.53 0.00 0.00 7.18 0.03 2.40 0.00 0.02 0.00 0.01 0.25 Q-11-02 5.22 2.78 2.52 0.01 0.00 7.15 0.04 2.44 0.00 0.00 0.00 0.00 0.25 Q-11-03 5.28 2.72 2.54 0.01 0.00 7.23 0.03 2.30 0.00 0.00 0.00 0.00 0.24 Q-11-04 5.19 2.81 2.57 0.00 0.00 6.97 0.03 2.56 0.00 0.01 0.01 0.02 0.27 Q-11-05 5.33 2.67 2.40 0.01 0.00 7.43 0.03 2.22 0.06 0.03 0.01 0.00 0.23 Q-11-06 5.24 2.76 2.49 0.00 0.00 7.21 0.04 2.41 0.00 0.01 0.00 0.01 0.25 Q-11-07 5.31 2.69 2.58 0.00 0.00 7.02 0.03 2.41 0.00 0.03 0.01 0.01 0.26 Q-11-08 5.23 2.77 2.60 0.00 0.00 6.96 0.04 2.49 0.00 0.00 0.00 0.01 0.26 Q-11-09 5.30 2.70 2.62 0.00 0.00 6.95 0.04 2.43 0.02 0.00 0.00 0.00 0.26 Q-11-10 4.42 3.58 2.61 0.00 0.00 6.87 0.04 2.22 0.01 0.00 0.00 0.00 0.24 Q-11-11 5.24 2.76 2.61 0.01 0.00 7.02 0.02 2.40 0.00 0.00 0.00 0.01 0.25 Q-11-12 5.18 2.82 2.59 0.01 0.00 7.07 0.03 2.43 0.00 0.01 0.00 0.01 0.26 Q-11-13 5.30 2.70 2.60 0.01 0.00 7.20 0.03 2.22 0.00 0.00 0.00 0.01 0.24 Q-11-14 5.18 2.82 2.54 0.01 0.00 7.19 0.03 2.37 0.00 0.03 0.00 0.02 0.25 Q-11-15 5.18 2.82 2.58 0.01 0.00 7.17 0.04 2.34 0.00 0.00 0.00 0.00 0.25 Q-13-01 5.61 2.39 2.55 0.00 0.09 3.18 0.07 5.98 0.00 0.00 0.00 0.00 0.65 Q-13-02 5.67 2.33 2.37 0.01 0.03 3.36 0.07 6.10 0.01 0.01 0.01 0.00 0.64 Q-13-03 5.63 2.37 2.48 0.01 0.06 3.22 0.07 6.06 0.00 0.01 0.00 0.00 0.65 Q-13-04 5.93 2.07 2.10 0.00 0.02 3.43 0.07 6.33 0.01 0.00 0.00 0.00 0.65 Q-13-05 5.76 2.24 2.31 0.01 0.04 2.97 0.07 6.54 0.00 0.02 0.02 0.00 0.69 Q-13-06 5.39 2.61 2.41 0.01 0.00 3.39 0.07 6.23 0.01 0.00 0.00 0.00 0.65 Q-13-07 5.54 2.46 2.39 0.00 0.00 3.38 0.07 6.29 0.00 0.02 0.00 0.00 0.65 Q-13-08 5.69 2.31 2.37 0.00 0.03 3.26 0.07 6.22 0.00 0.01 0.00 0.01 0.65 Q-13-09 5.50 2.50 2.47 0.01 0.00 3.33 0.07 6.12 0.00 0.01 0.00 0.00 0.65 Q-13-10 5.60 2.40 2.44 0.01 0.03 3.26 0.06 6.15 0.00 0.01 0.00 0.01 0.65 Q-13-11 5.69 2.31 2.47 0.01 0.10 3.20 0.07 6.03 0.00 0.00 0.00 0.00 0.65 Q-13-12 5.63 2.37 2.51 0.00 0.08 3.14 0.06 6.08 0.00 0.00 0.01 0.00 0.65 Q-13-13 5.67 2.33 2.49 0.01 0.09 3.16 0.07 6.05 0.00 0.00 0.01 0.00 0.65 Q-13-14 5.62 2.38 2.30 0.00 0.00 3.28 0.06 6.37 0.00 0.01 0.04 0.00 0.66 Q-13-15 5.58 2.42 2.45 0.00 0.02 3.29 0.07 6.14 0.00 0.00 0.01 0.00 0.65 Q-4-01 5.54 2.46 2.42 0.00 0.00 3.35 0.07 6.12 0.00 0.05 0.00 0.07 0.65 Q-4-02 6.24 1.76 2.20 0.02 0.15 3.05 0.06 5.98 0.01 0.00 0.01 0.01 0.66 Q-4-03 5.71 2.29 2.45 0.01 0.09 3.23 0.06 6.04 0.00 0.00 0.02 0.00 0.65 Q-4-04 5.64 2.36 2.42 0.00 0.04 3.31 0.07 6.09 0.01 0.02 0.00 0.01 0.65 Q-4-05 5.94 2.06 2.45 0.00 0.22 2.83 0.07 6.12 0.00 0.02 0.00 0.01 0.68 Q-4-06 5.64 2.36 2.38 0.01 0.02 3.33 0.07 6.17 0.00 0.00 0.00 0.01 0.65 Q-4-07 5.82 2.18 2.29 0.00 0.07 3.24 0.07 6.17 0.00 0.08 0.01 0.09 0.66 Appl. Sci. 2020, 10, 7444 17 of 24

Appl. Sci. 2020, 10, x FOR PEER REVIEW 17 of 24 Appl. Sci. 2020, 10, x FOR PEER REVIEW 17 of 24

Figure 18. Scatter diagrams of the wavelength and molecular formula of chlorite: (a) scatter diagram FigureFigure 18. 18.Scatter Scatter diagramsdiagrams of the wavelength wavelength and and molecular molecular formula formula of ofchlorite: chlorite: (a) (scattera) scatter diagram diagram of OH, Fe-OH, and Mg-OH bands and (b) scatter diagram of Mg, Fe, and AlVI cations; the data have ofof OH, OH, Fe-OH, Fe-OH,and and Mg-OHMg-OH bands and ( (bb)) scatter scatter diagram diagram of of Mg, Mg, Fe Fe,, and and Al AlVI VIcations;cations; the thedata data have have been normalized. beenbeen normalized. normalized.

FigureFigure 19. 19. Results Results of of the the absorption absorption depth depth of of chlorite chlorite: :( a(a) )mineral mineral map; map; ( b(b) )absorption absorption depth depth of of OH; OH; Figure 19. Results of the absorption depth of chlorite: (a) mineral map; (b) absorption depth of OH; (c()c )absorption absorption depth depth of of Fe Fe-OH;-OH; ( d(d) )absorption absorption depth depth of of Mg Mg-OH.-OH. (c) absorption depth of Fe-OH; (d) absorption depth of Mg-OH. Appl. Sci. 2020, 10, x FOR PEER REVIEW 18 of 24

6. Discussion

Appl.6.1. The Sci. 2020 Cause, 10 ,of 7444 Spectral Variation of Different Chlorites in Qidashan Iron Deposit 18 of 24 The spectral absorption features of chlorite are diagnostic to chemical composition variations. It can 6.be Discussioninferred that the different absorption wavelengths of chlorite in the wall rock of BIFs and high-grade iron ore are due to their different chemical compositions. Previous studies have shown that the 6.1. The Cause of Spectral Variation of Different Chlorites in Qidashan Iron Deposit wavelength of Fe-OH is positively correlated with the relative amount of Fe2+ and inversely with the Mg2+ in theThe octahedral spectral absorptioncation sites features[6,24,42,43]. of chloriteIn the Qidashan are diagnostic Iron Depos to chemicalit, the relationship composition between variations. the Itwavelength can be inferred positions that of the Fe di-OH,fferent Mg- absorptionOH, and OH wavelengths and the content of chlorite of cations in the in wallthe octahedral rock of BIFs layer and of high-gradechlorite shows iron that ore the are Fe due-OH to theirabsorption different band chemical is positively compositions. correlated Previous with the studies Fe and haveinversely shown related that thewith wavelength Mg and Mg of/(Mg Fe-OH + Fe) is positively(Figure 20 correlatedb,c), which with is consistent the relative with amount the result of Fe mentioned2+ and inversely above. with The theFe- MgOH2 + bandin the appears octahedral to be cation only sites weakly [6,24 influenced,42,43]. In the by Qidashan AlVI (Figure Iron 20 Deposit,a,d). The the other relationship absorption between bands themanifest wavelength similar positions characteristics of Fe-OH, to those Mg-OH, of Fe- andOH OH(Figure ands the21 and content 22). The of cations shorter in wavelength the octahedral of Fe layer-OH ofand/or chlorite Mg- showsOH indicates that the that Fe-OH chlorite absorption in wall rocks band around is positively BIFs is correlated Mg-rich (Q with-4 and the Q Fe-13), and whereas inversely the relatedlonger wavelength with Mg and is Mg indicative/(Mg + Fe)of Fe (Figure-rich chlorite 20b,c), whichin wall is rock consistent around with high the-grade result iron mentioned ore (Q-2 and above. Q- The11). Fe-OH band appears to be only weakly influenced by AlVI (Figure 20a,d). The other absorption bandsCompared manifest similarwith the characteristics absorption wavelength, to those of Fe-OHthe relationship (Figures 21 between and 22). the The absorption shorter wavelength depth and ofthe Fe-OH specific and chemical/or Mg-OH composition indicates that is not chlorite obvious. in wall The rocks absorption around depth BIFs is is Mg-rich usually (Q-4 related and Q-13), to the whereasabundance the of longer the absorber, wavelength but it is is indicative also strongly of Fe-rich affected chlorite by the ingrain wall size rock of aroundthe material high-grade [15]. Therefore, iron ore (Q-2the absorption and Q-11). depth of chlorite should be carefully applied in the practical mineral exploration.

FigureFigure 20.20. TheThe relationshiprelationship betweenbetween thethe wavelengthwavelength ofof Fe-OHFe-OH bandband andand thethe contentcontent ofof cationscations inin thethe octahedraloctahedral layerlayer ofof chlorite:chlorite: ((aa)) thethe relationshiprelationship betweenbetween Fe-OHFe-OH andand Fe;Fe; ((bb)) thethe relationshiprelationship betweenbetween Fe-OHFe-OH andand Mg;Mg; ((cc)) thethe relationshiprelationship betweenbetween Fe-OHFe-OH andand MgMg/(Mg+Fe);/(Mg+Fe); ((dd)) thethe relationshiprelationship betweenbetween Fe-OHFe-OH andand AlAlVIVI (the(the overlap points have been ooffsetﬀset forfor clarity).clarity). Appl. Sci. 2020, 10, 7444 19 of 24

Appl. Sci. 2020,, 10,, xx FORFOR PEERPEER REVIEWREVIEW 19 of 24

FigureFigure 21. 21.The The relationship relationshipip between betweenbetween the thethewavelength wavelengthwavelength ofof Mg-OHMgMg-OH band band and and the the content content of of cations cations in inthe the octahedraloctahedral layer layer of of chlorite: chlorite (:: a (()a) the the relationship relationship betweenbetween Mg Mg-OH-OH and and Fe; Fe; (b (b) )the the relationship relationship between between Mg-OHMg-OH and and Mg; Mg; (c )(c the) the relationship relationship between between Mg-OH Mg-OH andand MgMg/(Mg+Fe);/(Mg+Fe); (d (d) )the the relatio relationshipnship between between Mg-OH and AlVI (the overlap points have been offset for clarity). Mg-OH and AlVI (the overlap points have been oﬀset for clarity).

FigureFigure 22. 22.The The relationship relationship between between thethe wavelengthwavelength of OH OH band band and and the the content content of of cations cations in inthe the octahedraloctahedral layer layer of of chlorite: chlorite ((:: (a()a) the the relationship relationship betweenbetween OH and and Fe; Fe; ( (bb) )the the relationship relationship between between OH OH VI andand Mg; Mg; (c) ( thec) the relationship relationship between between OH OH and and Mg Mg/(Mg/(Mg + + Fe);Fe); (d) the the relationship relationship between between OH OH and and Al Al VI (the(the overlap overlap points points have have been been o ﬀoffsetset for for clarity)). clarity)).. Appl. Sci. 2020, 10, 7444 20 of 24

Compared with the absorption wavelength, the relationship between the absorption depth and the speciﬁc chemical composition is not obvious. The absorption depth is usually related to the abundance of the absorber, but it is also strongly aﬀected by the grain size of the material [15]. Therefore, the absorption depth of chlorite should be carefully applied in the practical mineral exploration.

6.2. Geological Significance and Application to Mineral Exploration The above spectra and EPMA analysis show that the chemical composition of chlorite around BIFs and high-grade iron ore is obviously different. This is probably due to their different formation conditions. Previous studies on the Qidashan Iron Deposit showed that the BIFs in this area were deposited in an oxygen-deficient environment on the seafloor and, finally, formed through low-to-medium-grade metamorphism. The high-grade iron ores were formed from BIFs protore through iron reactivation-reprecipitation process by neutral-acidic hydrothermal solution. Correspondingly, the chlorites related to BIFs were formed in the low-to-medium-grade regional metamorphism, whereas the chlorites associated with high-grade iron ores were formed by neutral-acidic Fe-rich hydrothermal alteration, and the protoliths is argillaceous siltstone [14,44–46]. The difference in the chemical composition of chlorite caused by the forming conditions is well recorded in the wavelength position of chlorite. The Gongchangling Iron Deposit is the largest high-grade iron ore deposit in Anshan-Benxi area, where, similarly, the wall rocks around the high-grade iron ore suffered strong chloritization and the chlorite in alteration wall rocks is Fe-rich [13,14,47]; this is also the case for other high-grade iron ore sites such as Nanfen, Wangjiapuzi, etc. [14]. In this means, the wavelength of chlorite can be used as a favorable tool for the high-grade iron ore exploration in the entire Anshan-Benxi area. Since the bandpass of hyperspectral imaging systems is less than 10 nm, these results may also provide a basis for the aerial or satellite hyperspectral remote sensing mineral exploration in the near future.

6.3. Advantages and Disadvantages of Laboratory-Based Hyperspectral Imaging Compared with point spectrometer, laboratory-based hyperspectral imaging has several advantages in extraction of mineralogical information, which are attributed to the high spatial resolution. First of all, the pixel size in hyperspectral image is much smaller than commonly used point spectrometer. For example, the probe size of the ASD FieldSpec-3 pro spectrometer used in this paper is 2 cm in diameter, whereas the pixel size of the HySpex image is 0.3 mm 0.3 mm. The relatively × small pixel size makes it possible to detect relatively small amounts of minerals whose spectral features are usually completely masked when measured by point spectrometer. The relatively small pixel size also provides spectra of pure minerals in hyperspectral imaging. Another advantage of hyperspectral imaging is that a large number of spectra are acquired simultaneously. By recording the number of pixels per mineral in the image, the mineralogy of samples can be quantified. This allows HySpex image to more accurately observe the details of minerals in the hand specimen, including mineral types, grain size, crosscutting, spatial distribution, paragenesis, etc., which can be used to determine the structure and texture of the rocks. The disadvantage of laboratory-based hyperspectral imaging is relatively low S/N compared to point spectrometer. If a better S/N is required, it will take a long time to collect the images. Even in the same sensor, the different mineral combinations have a certain impact on the S/N. If low-reflectivity minerals are accompanied by the minerals of high-reflectivity (such as quartz), the S/N of image is also higher and the spectra of minerals can be identified. If most of the minerals are dark with low-reflectivity, the low S/N makes it difficult to identify. Thompson also noted that it is difficult to identify minerals with a content of less than 5%, unless the sample is a mixture with quartz and the mineral is highly reflective. However, where low reflectance minerals are present, recognition may require 20% or more of the mineral in the sample [48]. Therefore, a smoothing method should be used in actual operations to improve the S/N of the hyperspectral images. Appl. Sci. 2020, 10, 7444 21 of 24

7. Conclusions Based on the study of hyperspectral imaging data in Qidashan iron mine, the following conclusions are drawn:

(1) The results of mineral mapping and chlorite spectral features extraction in Qidashan Iron Deposit show that chlorite both occurs in the wall rocks around BIFs and high-grade iron ore. The wavelength position of OH, Fe-OH, and Mg-OH absorptions of chlorite around high-grade iron ore are between 1400 and 1410, 2260 and 2265, and 2360 and 2370 nm, respectively, whereas the corresponding wavelengths are between 1390 and 1400, 2245 and 2255, and 2330 and 2340 nm around BIFs. (2) The higher spatial resolution laboratory-based hyperspectral imaging can reflect the specific information of mineral species, mineral content, structure, and even the polytypes of the rock sample. (3) The relationship between cations in the chlorite octahedral layer and the absorption wavelengths of OH, Fe-OH, and Mg-OH indicates that Fe is positively related to the wavelength, whereas Mg and Mg/(Mg + Fe) are inversely related. The wavelength appears to be weakly influenced by AlVI. Nevertheless, the relative absorption depths of the OH, Fe-OH, and Mg-OH are a poor guide to chemical composition of chlorite. (4) The difference in the chemical composition of chlorite caused by the different formation conditions of BIFs and high-grade iron ore in Qidashan Iron Deposit is well recorded in the wavelength position of chlorite. This difference of the wavelength, in turn, can be used as a favorable tool for the exploration of high-grade iron ore in the Anshan-Benxi area. This chlorite hyperspectral imaging tool may also provide a basis for the aerial or satellite hyperspectral remote sensing mineral exploration in Anshan-Benxi area.

Author Contributions: D.H. processed the data and wrote the paper; Y.Y. conducted the paper revision and the discussion of the results; J.F. provided financial support and useful suggestions for the paper; J.R.M. provided assistance in HySpex data acquisition; K.S. reviewed the paper. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the National Key Research and Development Program of China (grant 2016YFC0801603). Acknowledgments: We thank DiMap for access to the HySpex instruments used in the PSML, and we were also deeply indebted to the Qidashan Iron Deposit, Anshan Iron & Steel Group, for their kind help in field investigation and sampling. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Clark, R.N. Water frost and ice: The near-infrared spectral reflectance 0.65–2.5 µm. J. Geophys. Res. Space Phys. 1981, 86, 3087–3096. [CrossRef] 2. Hapke, B. Bidirectional reflectance spectroscopy: 1. Theory. J. Geophys. Res. Space Phys. 1981, 86, 3039–3054. [CrossRef] 3. Clark, R.N.; Roush, T.L. Reflectance spectroscopy: Quantitative analysis techniques for remote sensing applications. J. Geophys. Res. Space Phys. 1984, 89, 6329–6340. [CrossRef] 4. Bedini, E. Mineral mapping in the Kap Simpson complex, central East Greenland, using HyMap and ASTER remote sensing data. Adv. Space Res. 2011, 47, 60–73. [CrossRef] 5. Dalm, M.; Buxton, M.W.; Van Ruitenbeek, F.J.; Voncken, J.H. Application of near-infrared spectroscopy to sensor based sorting of a porphyry copper ore. Miner. Eng. 2014, 58, 7–16. [CrossRef] 6. Dalm, M.; Buxton, M.; Van Ruitenbeek, F. Discriminating ore and waste in a porphyry copper deposit using short-wavelength infrared (SWIR) hyperspectral imagery. Miner. Eng. 2017, 105, 10–18. [CrossRef] 7. Neal, L.C.; Wilkinson, J.J.; Mason, P.J.; Chang, Z. Spectral characteristics of propylitic alteration minerals as a vectoring tool for porphyry copper deposits. J. Geochem. Explor. 2018, 184, 179–198. [CrossRef] Appl. Sci. 2020, 10, 7444 22 of 24

8. De Linaje, V.A.; Khan, S.D.; Nicolás, V.A.D.L.D. Mapping of diagenetic processes in sandstones using imaging spectroscopy: A case study of the Utrillas Formation, Burgos, Spain. Sediment. Geol. 2017, 353, 114–124. [CrossRef] 9. Greenberger, R.N.; Ehlmann, B.L.; Jewell, P.W.; Birgenheier, L.P.; Green, R.O. Detection of organic-rich oil shales of the green river formation, Utah, with ground-based imaging spectroscopy. In Proceedings of the 2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, Los Angeles, CA, USA, 23 August 2016. 10. Zaini, N.; Van Der Meer, F.; Van Der Werff, H. Determination of Carbonate Rock Chemistry Using Laboratory-Based Hyperspectral Imagery. Remote Sens. 2014, 6, 4149–4172. [CrossRef] 11. Baissa, R.; Labbassi, K.; Launeau, P.; Gaudin, A.; Ouajhain, B. Using HySpex SWIR-320m hyperspectral data for the identification and mapping of minerals in hand specimens of carbonate rocks from the Ankloute Formation (Agadir Basin, Western Morocco). J. Afr. Earth Sci. 2011, 61, 1–9. [CrossRef] 12. Lampinen, H.M.; Laukamp, C.; Occhipinti, S.A.; Hardy, L. Mineral footprints of the Paleoproterozoic sediment-hosted Abra Pb-Zn-Cu-Au deposit Capricorn Orogen, Western Australia. Ore Geol. Rev. 2018, 104, 436–461. [CrossRef] 13. Wang, E.-D.; Xia, J.-M.; Fu, J.-F.; Jia, S.-S.; Men, Y.-K. Formation mechanism of Gongchangling high-grade magnetite deposit hosted in Archean BIF, Anshan-Benxi area, Northeastern China. Ore Geol. Rev. 2014, 57, 308–321. [CrossRef] 14. Li, H.; Yang, X.-Q.; Li, L.-X.; Zhang, Z.-C.; Liu, M.-J.; Yao, T.; Chen, J. Desilicification and iron activation–reprecipitation in the high-grade magnetite ores in BIFs of the Anshan-Benxi area, China: Evidence from geology, geochemistry and stable isotopic characteristics. J. Asian Earth Sci. 2015, 113, 998–1016. [CrossRef] 15. Clark, R.N. Spectroscopy of Rocks and Minerals, and Principles of Spectroscopy. In Remote Sensing for the Earth Sciences: Manual of Remote Sensing, 3rd ed.; Rencz, A.N., Ed.; John Wiley and Sons: New York, NY, USA, 1999; Volume 3, pp. 3–58. 16. Clark, R.N.; King, T.V.V.; Klejwa, M.; Swayze, G.A.; Vergo, N. High spectral resolution reflectance spectroscopy of minerals. J. Geophys. Res. Space Phys. 1990, 95, 12653–12680. [CrossRef] 17. King, T.V.V.; Clark, R.N. Spectral characteristics of chlorites and Mg-serpentines using high-resolution reflectance spectroscopy. J. Geophys. Res. Space Phys. 1989, 94, 13997–14008. [CrossRef] 18. Hunt, G.R. Spectral signatures of particulate minerals in the visible and near infrared. Geophysics 1977, 42, 501–513. [CrossRef] 19. Hunt, G.R. Near-infrared (1.3–2.4) µm spectra of alteration minerals—Potential for use in remote sensing. Geophysics 1979, 44, 1974–1986. [CrossRef] 20. Bishop, J.L.; Lane, M.D.; Dyar, M.D.; Brown, A.J. Reflectance and emission spectroscopy study of four groups of phyllosilicates: Smectites, kaolinite-serpentines, chlorites and micas. Clay Min. 2008, 43, 35–54. [CrossRef] 21. Bishop, J.L. Infrared Spectroscopic Analyses on the Nature of Water in Montmorillonite. Clays Clay Min. 1994, 42, 702–716. [CrossRef] 22. Yang, M.; Ye, M.; Han, H.; Ren, G.; Han, L.; Zhang, Z. Near-Infrared Spectroscopic Study of Chlorite Minerals. J. Spectrosc. 2018, 2018, 6958260. [CrossRef] 23. Burns, R.G. Mineralogical Applications of Crystal Field Theory; Cambridge University Press: London, UK, 1993; pp. 149–152. 24. Yang, K.; Lian, C.; Huntington, J.F.; Peng, Q.; Wang, Q. Infrared spectral reflectance characterization of the hydrothermal alteration at the Tuwu Cu–Au deposit, Xinjiang, China. Miner. Depos. 2005, 40, 324–336. [CrossRef] 25. James, H.L. Sedimentary facies of iron-formation. Econ. Geol. 1954, 49, 235–293. [CrossRef] 26. Zhai, M.; Santosh, M. Metallogeny of the North China Craton: Link with secular changes in the evolving Earth. Gondwana Res. 2013, 24, 275–297. [CrossRef] 27. Vaiphasa, C. Consideration of smoothing techniques for hyperspectral remote sensing. ISPRS J. Photogramm. Remote Sens. 2006, 60, 91–99. [CrossRef] 28. Fraser, S.J.; Whitbourn, L.; Yang, K.; Ramanaidou, E.; Connor, P.; Poropat, G.; Soole, P.; Mason, P.; Coward, D.; Phillips, R. Mineralogical Face-Mapping Using Hyperspectral Scanning for Mine Mapping and Control. In Proceedings of the 6th International Mining Geology Conference, Darwin, NT, Australia, 21 August 2006; pp. 227–232. Appl. Sci. 2020, 10, 7444 23 of 24

29. Ruffin, C.; King, R.L.; Younan, N.H. A Combined Derivative Spectroscopy and Savitzky-Golay Filtering Method for the Analysis of Hyperspectral Data. GISci. Remote Sens. 2008, 45, 1–15. [CrossRef] 30. Chen, J.; Jönsson, P.; Tamura, M.; Gu, Z.; Matsushita, B.; Eklundh, L. A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter. Remote Sens. Environ. 2004, 91, 332–344. [CrossRef] 31. Savitzky, A.; Golay, M.J.E. Smoothing and Differentiation of Data by Simplified Least Squares Procedures. Anal. Chem. 1964, 36, 1627–1639. [CrossRef] 32. Ostertagov, E.; Ostertag, O. Methodology and application of Savitzky-Golay moving average polynomial smoother. Glob. J. Pure Appl. Math. 2016, 12, 3201–3210. 33. Tong, D.; Jia, Z.; Qin, X.; Cao, C.; Chang, C. An Optimal Savitzky-golay Filtering Based Vertical Handoff Algorithm in Heterogeneous Wireless Networks. J. Comput. 2014, 9, 2685–2690. [CrossRef] 34. Mao, Y.; Wang, D.; Liu, S.; Song, L.; Wang, Y.; Zhao, Z. Research and Verification of a Remote Sensing BIF Model Based on Spectral Reflectance Characteristics. J. Indian Soc. Remote Sens. 2019, 47, 1051–1061. [CrossRef] 35. Kruse, F.A.; Bedell, R.L.; Taranik, J.V.; Peppin, W.A.; Weatherbee, O.; Calvin, W.M. Mapping alteration minerals at prospect, outcrop and drill core scales using imaging spectrometry. Int. J. Remote Sens. 2011, 33, 1780–1798. [CrossRef][PubMed] 36. Boardman, J.W. Analysis, understanding and visualization of hyperspectral data last convex sets in n-space. In Proceedings of the SPIE Defense and Security-The International Scoiety for Optical Engineering, Orlando, FL, USA, 17 April 1995. 37. Kruse, F.; Lefkoff, A.; Boardman, J.; Heidebrecht, K.; Shapiro, A.; Barloon, P.; Goetz, A. The spectral image processing system (SIPS)—Interactive visualization and analysis of imaging spectrometer data. Remote Sens. Environ. 1993, 44, 145–163. [CrossRef] 38. Duke, E.F. Near infrared spectra of muscovite, Tschermak substitution, and metamorphic reaction progress: Implications for remote sensing. Geology 1994, 22, 621–624. [CrossRef] 39. Van Der Meer, F. Spectral reflectance of carbonate mineral mixtures and bidirectional reflectance theory: Quantitative analysis techniques for application in remote sensing. Remote Sens. Rev. 1995, 13, 67–94. [CrossRef] 40. Kurz, T.H.; Dewit, J.; Buckley, S.J.; Thurmond, J.B.; Hunt, D.W.; Swennen, R. Hyperspectral image analysis of different carbonate lithologies (limestone, karst and hydrothermal dolomites): The Pozalagua Quarry case study (Cantabria, North-west Spain). Sedimentology 2011, 59, 623–645. [CrossRef] 41. Zaini, N.; Van Der Meer, F.D.; Van Der Werff, H. Effect of Grain Size and Mineral Mixing on Carbonate Absorption Features in the SWIR and TIR Wavelength Regions. Remote Sens. 2012, 4, 987–1003. [CrossRef] 42. Yang, K.; Huntington, J.F.; Browne, P.R.; Ma, C. An infrared spectral reflectance study of hydrothermal alteration minerals from the Te Mihi sector of the Wairakei geothermal system, New Zealand. Geothermics 2000, 29, 377–392. [CrossRef] 43. McLeod, R.L.; Gabell, A.R.; Gardavsky, V. Chlorite infrared spectral data as proximity indicators of volcanogenic massive sulfide mineralisation. In Proceedings of the Pacific Rim Congress 87, Gold Coast, Australia, 1 January 1987; pp. 321–324. 44. Li, H.; Zhang, Z.; Li, L.; Zhang, Z.; Chen, J.; Yao, T. Types and general characteristics of the BIF-related iron deposits in China. Ore Geol. Rev. 2014, 57, 264–287. [CrossRef] 45. Liu, M.J.; Zeng, Q.D.; Li, H.M.; Li, L.X.; Wen, Y.; Yao, L.D.; Gao, Y.S. Metallogenic epoch of high-grade iron ore deposits of iron activation and enrichment genesis in Anshan-Benxi area of Liaoning: Re-Os isotopic dating evidence of molybdenite from Qidashan iron deposit. Miner. Depos. 2017, 36, 237–249. 46. Li, Y.H.; Zhang, Z.J.; Hou, K.J.; Duan, C.; Wang, D.F.; Hu, G.Y. The genesis of Gongchangling high-grade iron ores, Anshan-Benxi area, Liaoning province, NE China: Evidence from Fe-Si-O-S isotopes. Acta Geol. Sin. 2014, 88, 2351–2372. Appl. Sci. 2020, 10, 7444 24 of 24

47. Chen, G.Y.; Sun, D.S.; Sun, C.M.; Li, M.H.; Wang, X.F.; Wang, Z.F. Genesis of Gongchanglin iron deposit. J. Mineral. Petrol. 1984, 4, 1–226. 48. Thompson, A.J.B.; Hauﬀ, P.L.; Robitaille, A.J. Alteration mapping in exploration: Application of short-wave infrared (SWIR) spectroscopy. Soc. Econ. Geol. Newsl. 1999, 39, 15–27.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).