geosciences

Article Analysis by UAV Digital of Folds and Related Fractures in the Monte Antola Flysch Formation (Ponte Organasco, Italy)

Niccolò Menegoni * ID , Claudia Meisina, Cesare Perotti and Matteo Crozi

Department of Earth and Environmental Sciences, University of Pavia, via A. Ferrata 1, I-27100 Pavia, Italy; [email protected] (C.M.); [email protected] (C.P.); [email protected] (M.C.) * Correspondence: [email protected]; Tel.: +39-0382-985849

 Received: 25 July 2018; Accepted: 7 August 2018; Published: 9 August 2018 

Abstract: The deformation structures (folds and fractures) affecting Monte Antola flysch formation in the area of Ponte Organasco (Northern Apennines-Italy) were analyzed by Unmanned Aerial Vehicle Digital Photogrammetry (UAVDP). This technique allowed the realization of Digital Outcrop Models (DOMs) interpreted in a stereoscopic environment by collecting a large number of digital structural measures (strata, fractures and successively fold axes and axial planes). In particular, by UAVDP was possible to analyze the relationships between folds and fractures all along the study structures. The structural analysis revealed the presence of a series of NE-vergent folds characterized by a typical Apenninic trend and affected by four main sets of fractures. Fractures are always sub-orthogonal to the bedding, maintains constant angular relationships with the bedding and seems linked to the folding deformation. The study shows that the UAVDP technique can overcome the main limitations of field structural analysis such as the scarce presence and the inaccessibility (total or partial) of rock outcrops and allows for acquiring images of rock outcrops at a detailed scale from user-inaccessible positions and different points of view and analyze inaccessible parts of outcrops.

Keywords: structural geology; Unmanned Aerial Vehicle; digital photogrammetry;

1. Introduction In the last twenty years, the applications of remote sensing investigations for the construction of Digital Outcrop Models (DOMs) have rapidly increased in geosciences e.g., [1–11]. The most common techniques that are used to generate high detailed DOMs are Laser Scanner (LS) and Digital Photogrammetry (DP). Whereas LS could be very expensive and complex as for survey planning (heavy and bulky equipment), DP allows for obtaining high resolution data with a lower cost and a more user-friendly survey planning [7,12]. Developments in RGB cameras and Unmanned Aerial Vehicle (UAV) technologies [13] are quickly increasing the applications for UAV-based Digital Photogrammetry (UAVDP) in geosciences e.g., [7,9,14–18]. The field structural analysis could be often hampered by total or partial inaccessibility of rock outcrops [5] and by their unfavorable orientations [11] that do not permit acquiring an appreciable amount of structural data. Moreover, compass and clinometer field measurements of discontinuities could be affected by some biases (e.g., orientation and truncation biases) due to sampling techniques (e.g., scanline, window, etc.) or to the local variation in the orientation of measured features (e.g., waved/undulated surfaces). Sturzenegger and Stead [5] and Sturzenegger et al. [19] show how Digital Photogrammetry (DP) could help obtain more accurate parameters of position, orientation, length and curvature of discontinuity features. Tavani et al. [11] show how DP could help overcome limitations of structure exposure, giving the possibility to export orthorectified images of DOMs that

Geosciences 2018, 8, 299; doi:10.3390/geosciences8080299 www.mdpi.com/journal/geosciences Geosciences 2018, 8, 299x FOR PEER REVIEW 22 o of f 16 16 limitations of structure exposure, giving the possibility to export orthorectified images of DOMs that well fit fit the orientation orientation ofof geological geological structures. structures. Nevertheless, Terrestrial Digital Photogrammetry (TDP) often often suffers suffers limitations, limitations, such such as as occlusion occlusion and and vertical vertical orientation orientation biases biases of ofthe the highest highest part part of theof the outcrop outcrop [5], [5 due], due to tothe the lack lack of of an an optimal optimal positioning positioning of of the the camera camera wit withh respect respect to to the the rock exposure. These These limitations limitations can be be overcome overcome by by UAVDP UAVDP possibility to acquire images of an outcrop fromfrom user user-inaccessible-inaccessible positionspositions [[9].9]. Our stud studyy aims to demonstrate how the application of UAVDP UAVDP can dramatically improveimprove the structural analysis of folded folded and and fractured fractured rocks rocks by by acquiring a great amount of digital data from inaccessibleinaccessible (partially or totally) outcrops outcrops and increasing increasing the the sampling sampling of of t thehe deformation deformation structures. structures. InIn particular, thisthis studystudy demonstratesdemonstrates as UAVDP technique allows for analyzing the relationships between folds and fractures in a flyschflysch formation.formation. InIn thethe North-WesternNorth-Western Italian Italian Appennines, Appennines, the the Ponte Ponte Organasco Organasco area area offers offers the possibility the possibility to study to studythe meso-scale the meso deformations-scale deformations that affect that Monteaffect Monte Antola Antola formation formation (Figure (Figure1). In this 1).area In this the area erosion the erosionof the Trebbia of the Trebbia river creates river several creates spectacular several spectacular cliffs, almost cliffs, completely almost completely inaccessible inaccessible by field surveys,by field surveys,and characterized and characterize by severald by folds several and folds brittle and structures. brittle structures.

Figure 1. 1. TectonicTectonic sketch sketch of North of North-Western-Western Apennines: Apennines: 1: Quaternary 1: Quaternary Deposits; Deposits; 2: Te rtiary 2: Pie Tertiary dmont BasinPiedmont de posits; Basin 3: deposits;Epiligurian 3: Succession; Epiligurian 4 Succession;: Inte rnal Ligurian 4: Internal Units; Ligurian 5: Voltri Units; Group; 5: 6:Voltri Antola Group; Unit; 7:6: Extern Antolaal Unit; Ligurian 7: External Units; 8: LigurianSubligurian Units; Units; 8: Subligurian9: Tuscan Units Units; (Tuscan 9: Tuscan Nappe); Units 10: (TuscanSestri-Voltaggio Nappe); Zone.10: Sestri-Voltaggio Zone.

2. Geological Geological Setting The study study area is located in NW Apennines, near Ponte Organasco village at the confluenceconfluence between AvagnoneAvagnone and and Trebbia Trebbia rivers, rivers, nearly nearly 100 100 km km north north of Genoa of Genoa (Figure (Figure2). In this2). In area, this the area, Antola the AntolaUnit (AU) Unit crops (AU) cropsout extensively out extensively and isand affected is affected by aby great a great number number of of folds, folds, faults faultsand and fractures developed in absence of metamorphism.

Geosciences 2018, 8, 299 3 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 3 o f 16

Figure 2. Ge ological map of the study are a. Figure 2. Geological map of the study area. From bottom to the top, the AU unit [20–22] is composed by:

From• varicolored bottom to hemipelagic the top, the AUshales unit (the [20 –Late22] isCenomanian composed- by:Early Turonian Montoggio Shale- • varicolored“Argilliti hemipelagic di Montoggio”) shales followed (the Late by thin Cenomanian-Early-bedded, siliciclastic Turonian and carbonatic Montoggio turbidites Shale-“Argilliti (Early Campanian Gorreto Sandstone—“Arenarie di Gorreto”); di Montoggio”) followed by thin-bedded, siliciclastic and carbonatic turbidites (Early Campanian • carbonate turbidite (Early Campanian-Early Maastrichtian Monte Antola Formation- Gorreto Sandstone—“Arenarie di Gorreto”); “Formazione di Monte Antola” or “Calcare del Monte Antola”); • carbonate• siliciclastic turbidite-carbonatic (Early Campanian-Earlyturbidite (Early Maastrichtian Maastrichtian-lower Monte Late Antola Maastricht Formation-“Formazioneian Bruggi- di MonteSelvapiana Antola” Formation or “Calcare—“Formazione del Monte di Antola”); Bruggi-Selvapiana” and the Late Maastrichtian-Late • siliciclastic-carbonaticPaleocene Pagliaro turbiditeShale—“Argilliti (Early di Maastrichtian-lower Pagliaro”). Late Maastrichtian Bruggi-Selvapiana Formation—“FormazioneAU is topped by the Tertiary di Bruggi-Selvapiana” Piedmont Basin (TPB) and, a wedge the Late‐top‐ Maastrichtian-Latebasin located on top of Paleocene the Pagliarojunction between Shale—“Argilliti the Western di Pagliaro”). Alps and the Northern Apennines, that is characterized by a continuous deposition from the Mid Eocene (Monte Piano Marls) to the Late Miocene [23]. AU is topped by the Tertiary Piedmont Basin (TPB), a wedge-top-basin located on top of the The examined outcrops are all in the Monte Antola Formation. Locally, it is composed of junctiondominantly between turbiditic the Western calcareous Alps and graded the Northern beds, calcareous Apennines, sandstones, that is sandstones characterized and by marlstones, a continuous depositionwith rare from intervening the Mid Eocene thin, carbonate (Monte Piano-free graded Marls) layers to the Lateor uniformly Miocene fine [23-].grained shale. The Theformation examined is interpreted outcrops as a are deep all-sea in basin the Monte plain deposit, Antola with Formation. local latera Locally,l facies variations it is composed which of dominantlyrange from turbiditic proximal, calcareous thick-bedded graded turbidites beds, calcareousto distal turbidites sandstones, with predominantly sandstones and thickening marlstones, with rareupward intervening cycles and thin, a high carbonate-free percentage of gradedpelites. The layers thickness or uniformly of the more fine-grained competent shale. beds varies, The formation but is interpretedon average as it aranges deep-sea between basin 1.5 plain and 2 deposit, m. with local lateral facies variations which range from proximal,In thick-bedded the study area, turbidites the AU (locally to distal represented turbidites by Monte with predominantlyAntola Formation thickening and Montoggio upward Shale) cycles overthrusts onto the “Argille a palombini di Monte Veri” Formation, belonging to the External and a high percentage of pelites. The thickness of the more competent beds varies, but on average it Ligurian Unit (Figure 2). ranges between 1.5 and 2 m. The collisional belt of the Northern Apennines can in fact be subdivided into three major groups Inof unitsthe study characterized area, the by AU defined (locally structural represented and paleogeographic by Monte Antola domains: Formation (1) Tuscan and Units Montoggio; (2) Sub Shale)- overthrusts onto the “Argille a palombini di Monte Veri” Formation, belonging to the External Ligurian Unit (Figure2). The collisional belt of the Northern Apennines can in fact be subdivided into three major groups of units characterized by defined structural and paleogeographic domains: (1) Tuscan Units; Geosciences 2018, 8, 299 4 of 16

(2) Sub-Ligurian Units and (3) Ligurian Units [24]. The NE sector of the Northern Apennines is mainly representedGeosciences by 2018 the, 8, x Ligurian FOR PEER UnitsREVIEW (Figure 1). 4 o f 16 Ligurian Units are placed at top of the nappe pile due to the Apenninic tectonic stage (OligoceneLigurian and Units Miocene-Pleistocene) and (3) Ligurian Units thrusting [24]. The onto NE the sector more of external the Northern domain Apennines of continental is mainl marginy of Adriarepresented (Sub-Ligurian by the Ligurian and Tuscan Units domain;(Figure 1). [ 20 ]). These units are characterized by an assemblage of ophioliticLigurian sequences Units of are Jurassic placed age at top and of thethe nappe related pile sedimentary due to the Apenninic successions-(Late tectonic stage Jurassic (Oligocene to Middle Eoceneand [ 25Miocene]) and-Pleistocene) represent the thrusting transition onto betweenthe more theexternal Ligurian-Piedmont domain of continental ocean margin fragments of Adria and the (Sub-Ligurian and Tuscan domain; [20]). These units are characterized by an assemblage of ophiolitic Adria continental margin [22]. sequences of Jurassic age and the related sedimentary successions-(Late Jurassic to Middle Eocene Elter [24] subdivided Ligurian units into two groups according to their structural and stratigraphic [25]) and represent the transition between the Ligurian-Piedmont ocean fragments and the Adria features:continental Internal margin Ligurian [22]. Units (ILU) and External Ligurian Unit (ELU). The ILU derived from oceanic basinElter setting[24] subdivided whereas theLigurian ELU fromunits the into transition two groups from according ocean basin to totheir continental structural margin and of the Adriaticstratigraphic plate features: [25]. Internal Ligurian Units (ILU) and External Ligurian Unit (ELU). The ILU Thederived ILU from arethrusted oceanic basin onto thesetting ELU whereas along the the Levanto-Ottone ELU from the transition tectonic line from (Figure ocean1 ).basin to Thecontinental AU is margin geometrically of the Adriatic the highest plate [25]. unit outcropping in this sector of the Apenninic chain, neverthelessThe ILU it has are been thrusted recently onto the ascribed ELU along to eastern the Levanto ELU-Ottone due to tectonic the lack line of (Figure ophiolitic 1). debris in its basal complexThe AU [22 is]. geometrically the highest unit outcropping in this sector of the Apenninic chain, nevertheless it has been recently ascribed to eastern ELU due to the lack of ophiolitic debris in its The strong deformations affecting the Monte Antola Formation are difficult to investigate because basal complex [22]. of the inaccessibility of the outcrops, often represented by vertical or sub-vertical cliffs or slopes. The strong deformations affecting the Monte Antola Formation are difficult to investigate Moreover,because the of ductile the inaccessibility deformations of the are accompaniedoutcrops, often byrepresented well-developed by vertical fracture or sub networks-vertical thatcliffs enhance or circulationslopes. ofMoreover, fluid causing the ductile frequent deformations problems are of accompanied slope stability. by well-developed fracture networks that enhance circulation of fluid causing frequent problems of slope stability. 3. Study Area 3. Study Area The study area is located near Ponte Organasco village (Lat. 44.685◦ N; Long. 9.306◦ E), where the AU overthrustsThe study on thearea ELU is located locally near represented Ponte Organasco by the village “Argille (Lat. a palombini 44.685° N; diLong. Monte 9.306° Veri” E), where formation, that isthe made AU upoverthrusts by heavily on deformed the ELU locally dark shales represented interbedded by the with“Argille silica-rich a palombini limestones, di Monte and Veri” polygenic breccias,formation, ophiolites that is and mad olistolites.e up by heavily Due todeformed the proximity dark shales of this interbedded thrust, the with Monte silica Antola-rich limestones, Formation is and polygenic breccias, ophiolites and olistolites. Due to the proximity of this thrust, the Monte affected by a series of thrust related folds, faults and fractures (Figure2). Antola Formation is affected by a series of thrust related folds, faults and fractures (Figure 2). Two outcrops of the Monte Antola Flysch, both located along Trebbia river (Figure3), have been Two outcrops of the Monte Antola Flysch, both located along Trebbia river (Figure 3), have been studiedstudied by means by means of DP. of DP.

Figure 3. The study outcrops marked on (a) an orthorectified aerial image and (b) a perspective image. Figure 3. a b In (b)The the positions study outcrops of the Ground marked Control on ( ) Points an orthorectified (GCPs) are shown. aerial image and ( ) a perspective image. In (b) the positions of the Ground Control Points (GCPs) are shown. The first outcrop (O1) is a sub-vertical cliff 100 m high and more than 200 m wide, oriented SW- TheNE, showing first outcrop well exposed (O1) is folding a sub-vertical structures cliff (Figure 100 4a). m high The second and more outcrop than (O2) 200 is ma sub wide,-vertical oriented SW-NE,cliff showing40 m high well and exposed 100 m wide, folding oriented structures SW-NE (Figure (Figure4a). 4b). The These second outcrops outcrop were (O2) investigated is a sub-vertical by UAVDP because of the impossibility to obtain satisfactory information by structural field surveys. cliff 40 m high and 100 m wide, oriented SW-NE (Figure4b). These outcrops were investigated by UAVDP because of the impossibility to obtain satisfactory information by structural field surveys.

Geosciences 2018, 8, 299 5 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 5 o f 16

Figure 4. Orthorectified images of Digital Outcrop Models DOMs representing outcrops 1 (a) and Figure 4. Orthore ctifie d images of Digital Outcrop Mode ls DOMs representing outcrops 1 (a) and 2 2 (b).(b The). The view vie w points points of the the orthorectified orthorectified images images are horizontal are horizontal and sub and-orthogonal sub-orthogonal to the mean to theslope mean slopedirection. direction. The The position position of the of theGCPs GCPs and the and control the control planes planes (natural (natural surface surfacess whose attitude whose was attitude was measuredme asured in in the the fie field ld and and on onthe the DOMs) DOMs) are indicated are indicated with ye with llow yellow and re d and dots, red re spectively. dots, respectively. The The positionposition of GCP 2 2 is is shown shown in ina Unmanne a Unmanned d Ae rial Aerial Ve hicle Vehicle (UAV) (UAV) (c) and (c )fie and ld (d field) image (d)s. images.

4. Methodology4. Methodology

4.1. UAV4.1. UAV Digital Digital Photogrammetry Photogrammetry (UAVDP) (UAVDP) SurveysSurveys Due to the inaccessibility and huge dimension of the outcrops (sub-vertical slopes more than 50 Due to the inaccessibility and huge dimension of the outcrops (sub-vertical slopes more than 50 m m in height) two UAVDP surveys were conducted in order to obtain two high-resolution DOMs and in height) two UAVDP surveys were conducted in order to obtain two high-resolution DOMs and to avoid TDP problems of censoring and bias induced by unfavorable orientations of the camera [5]. to avoidA mirrorless TDP problems Sony of ILCE censoring-6000 24 and Megapixel bias induced camera by (APS unfavorable-C sensor), orientations (Sony Europe of the Limited, camera [5]. A mirrorlessWeybridge, Sony UK) ILCE-6000 with 16 mm 24 lens, Megapixel was mounted camera on (APS-C a hexacopter sensor), (UAV) (Sony using Europe a 2-axis Limited, gimbal. Weybridge, Main UK) withspecifications 16 mm lens, of UAV was and mounted camera onare areported hexacopter in Table (UAV) 1. using a 2-axis gimbal. Main specifications of UAV and camera are reported in Table1. Table 1. UAV and UAV-came ra spe cifications.

Table 1. UAV and UAV-camera specifications.Empty UAV-Camera UAV Type Diameter Engines Rotor Diameter Weight Weight UAVhexacopter Type Diameter115 cm Enginesbrushless Rotor15 inch Diameter (38.1 cm) Empty6.9 Weight kg UAV-Camera8.3 kg Weight hexacopter 115Sensor cm brushless 15 inch (38.1 cm) 6.9 kg 8.3 kg Camera Sensor Size Image Size Pixel Size Focal Length Camera SensorT Typeype Sensor Size Image Size Pixel Size Focal Length SonySony 6000 ILCE6000 ILCE APS-CAPS-C 23.523.5× ×15.6 15.6 mm mm 60006000× ×4000 4000 pixels pixels 3.923.92× ×3.90 3.90µ mµm 16 16 mm mm

The UAVDP surveys were conducted with an oblique orientation of the on-board camera and The UAVDP surveys were conducted with an oblique orientation of the on-board camera and the acquisition of digital photographs with a minimum overlap and sidelap of about 85% and 65%, the acquisitionrespectively. of In digital order to photographs capture the complex with a minimumgeometry of overlap the outcrops and sidelapand to improve of about the 85% accuracy and 65%, respectively.of the generated In order DOMs, to capture the images the complex were acquired geometry from of positions the outcrops parallel and (strips to improve of photographs the accuracy of thetaken generated along a fly DOMs, line) and the convergent images were to the acquired outcrop [26]. from The positions distance parallelfrom the camera (strips to of the photographs closest takenrock along surface a fly line)was from and convergent15 to 52 m. toAccording the outcrop to Birch [26]. [26], The giving distance to the from camera the camera setting to and the the closest rockcamera surface-outcrop was from distances 15 to the 52 m.Ground According Sample Distances to Birch [(GSDs)26], giving of the tophotographs the camera are setting between and 4 the camera-outcropand 13 mm. distancesThe flights the were Ground flown Sampleunder manual Distances control (GSDs) in a sequence of the photographs of back-and- areforward between flight 4 and 13 mm.lines The to flightscover the were full flownvertical under extent manual of the rock control outcrops in a (Figure sequence 5). of back-and-forward flight lines to cover the full vertical extent of the rock outcrops (Figure5). Geosciences 2018, 8, 299 6 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 6 o f 16

Figure 5. (a) Front and (b) top views of the flight plan showing flight line s and dire ction of flight used Figure 5. (a) Front and (b) top views of the flight plan showing flight lines and direction of flight used for the photogrammetric acquisition of DOM1. Picture s we re acquired from positions paralle l (red for theline photogrammetric s) and conve rgent (yellow acquisition line s) ofto the DOM1. outcrop. Pictures The second were UAV acquired survey from conducted positions with the parallel same (red lines)flight and convergentsche me. (yellow lines) to the outcrop. The second UAV survey conducted with the same flight scheme. All photographs (334 for DOM1 and 145 for DOM2) were georeferenced and oriented by the AllUAV photographs-on board GNSS/IMU (334 for DOM1instrumentations and 145 forand DOM2) 11 natural were Ground georeferenced Control Points and oriented(GCPs—see by the Figures 3 and 4) that were acquired using a Topcon Hyper Pro Real Time Kinematic (RTK) differential UAV-on board GNSS/IMU instrumentations and 11 natural Ground Control Points (GCPs—see GPS system, with a horizontal mean error of 1.7 cm and a vertical mean error of 3.2 cm. Figures3 and4) that were acquired using a Topcon Hyper Pro Real Time Kinematic (RTK) differential GPS system,4.2. Digital with Outcrop a horizontal Models Development mean error and of A 1.7nalysis cm and a vertical mean error of 3.2 cm.

4.2. DigitalThe Outcrop images Models were processed Development in two and Analysisdifferent blocks using the (SfM) technique [7,27,28], using the software package Agisoft Photoscan (version 1.2.5, Agisoft LLC, St. ThePetersburg, images Russia) were [29]. processed The alignment in two of differentthe images blocks and theusing development the Structure of the sparse from point Motion clouds (SfM) techniquewere conducted [7,27,28], using using 11 the non software-collinear GCPs package (6 and Agisoft 5 GCPs Photoscan respectively (version for the two 1.2.5, outcrops, Agisoft see LLC, St. Petersburg,Figures 3 and Russia) 4) and [29the]. full The resolution alignment of the of photographs. the images and The thedense development point clouds were of the calculated sparse point cloudssubsampling were conducted the photographs using 11 non-collinearby a factor of 2 in GCPs each dimension (6 and 5 GCPs in order respectively to develop foreasy the-manageable two outcrops, see Figurespoint clouds3 and4 (<)51 and millions the full of resolution points) in a of reasonable the photographs. time. Finally The two dense textured point meshes clouds were created calculated (mean surface of the triangle faces ~1 cm2). The accuracy of the developed DOMs, will be discussed subsampling the photographs by a factor of 2 in each dimension in order to develop easy-manageable in the next paragraph. DOM1 and DOM2 are respectively referred to O1 and O2. Tables 2 and 3 show point clouds (<51 millions of points) in a reasonable time. Finally two textured meshes were created the main features of the DOMs. (mean surface of the triangle faces ~1 cm2). The accuracy of the developed DOMs, will be discussed in the next paragraph. DOM1 and DOM2Table are 2. respectively Characte ristics referred of DOM1. to O1 and O2. Tables2 and3 show the main features of the DOMs. Number of Images 334 Surface of the Model (m2) 4513 Table 2. Characteristics of DOM1. Dist ance Camera-Outcrop Range (m) 31–52 Mean Distance Camera-Outcrop (m) 39 Number of Images 334 Mean GSD at the Outcrop Surface (mm) 9.5 Surface of the Model (m2) 4513 DistanceNumber Camera-Outcrop of Dense Cloud Points Range (m)~50.5 31–52 million MeanNumber Distance of Triangular Camera-Outcrop Facets (m)~10.5 39 million Mean GSD at the Outcrop Surface (mm) 9.5 Number of Dense Cloud Points ~50.5 million Number of Triangular Facets ~10.5 million Geosciences 2018, 8, 299 7 of 16

Table 3. Characteristics of DOM2.

Number of Images. 145 Surface of the Model (m2) 224 Distance Camera-Outcrop Range (m) 15–20 Mean Distance Camera-Outcrop (m) 3.9 Mean GSD at the Outcrop Surface (mm) 47 ~31 Number of Dense Cloud Points million Number of Triangular Facets ~3 million

The DOMs were interpreted using the open-source software Cloud Compare version 2.9 [30], that allows the stereo-vision of the models and digital sampling of bedding and fractures. The stereo-visualization and structural interpretation of the DOMs were performed by a Planar SD2220W stereoscopic monitors. Only in stereo-vision, in fact, it is possible to understand the real nature and geometry of the structures (strata, fractures, lineations, folds, etc.) and to avoid misinterpretations due to 2D visualization on standard monitors of 3D objects. The procedure to measure the orientation of a bedding or fracture plane using Cloud Compare software consists of point-picking the intersection between the selected fracture/bedding surface and the DOM’s surface. Then the best-fit plane of the picked points is calculated. Bedding and fracture orientation were also mapped using the compass [31] new plugin of Cloud Compare that permits to sample traces on point clouds in a semi-automatic way and to extract their position, orientation and length.

4.3. Accuracy of DOMs The accuracy of a DOM can be distinguished in absolute and relative accuracy [32]. Absolute accuracy refers to the difference between the reconstructed position of a point of the DOM, calculated by SfM procedure, and the location of the same point on Earth (e.g., latitude, longitude, and elevation). Relative accuracy is a measure of positional coherence between a point and the neighboring points that is the difference in length and azimuth of the vectors that join two points on Earth, and the corresponding length and azimuth of the same vectors in the model. For geological and structural outcrop studies, relative accuracy is certainly more important because the measurements (e.g., attitudes of plane and surfaces) calculated in a DOM characterized by good relative accuracy are equivalent to measurements made on the outcrop. An accurate absolute location of these measures is often less important for geological purposes. The influence of georeferenced UAVDP models on surface orientation measurements (e.g., beds and fractures dip and dip azimuth) are comparable to the accepted range of errors of the user/instrument orientation measurements [33]. This range (ca. ±2◦) could be defined by the accuracy of high level compass-clinometer instrument [34] and by the statistical significance of field measurements of not perfectly planar surfaces (e.g., slightly undulated surfaces). The absolute accuracy of the study DOMs georeferenced using 11 GCPs was calculated by comparing GCPs coordinate measured by the RTK differential GPS system and those estimated in the model (Table4). Their comparison shows a mean horizontal difference of 0.41 m with a standard deviation of 0.16 m and a vertical difference of 0.30 m with a standard deviation of 0.18 m.

Table 4. Absolute accuracy (GCPs errors) of DOMs in meters.

GCPs and CKPs Horizontal Error Vertical Error Total Error 1 (GCP) 0.243 0.203 0.316 2 (CKP) 0.519 0.523 0.737 3 (GCP) 0.357 0.224 0.421 4 (GCP) 0.389 0.113 0.405 5 (CKP) 0.639 0.274 0.695 Geosciences 2018, 8, 299 8 of 16

Table 4. Cont.

GCPs and CKPs Horizontal Error Vertical Error Total Error

Geosciences 20186, (GCP) 8, x FOR PEER REVIEW 0.190 0.209 0.282 8 o f 16 7 (GCP) 0.369 0.272 0.458 8 (CKP)6 (GCP) 0.4120.190 0.209 0.253 0.282 0.483 9 (GCP)7 (GCP) 0.1120.369 0.272 0.048 0.458 0.122 10 (CKP)8 (CKP) 0.2910.412 0.253 0.234 0.483 0.373 11 (GCP) 0.351 0.376 0.515 9 (GCP) 0.112 0.048 0.122 DOM1 Mean Error 0.389 0.258 0.476 10 (CKP) 0.291 0.234 0.373 DOM2 Mean Error 0.307 0.237 0.390 GCPs Mean11 (GCP) Error 0.2870.351 0.376 0.206 0.515 0.360 CKPs MeanDOM1 Error Mean Error0.465 0.389 0.258 0.321 0.476 0.572 DOM2 Mean Error 0.307 0.237 0.390 GCPs Mean Error 0.287 0.206 0.360 The relative accuracyCKPs ofMean the Error model was0.465 evaluated comparing:0.321 0.572 1. the lengths and orientations of the vectors that join the GCPs estimated both by the GPS The relative accuracy of the model was evaluated comparing: measurements and by the digital measurements performed on the model (Table5). 1. the lengths and orientations of the vectors that join the GCPs estimated both by the GPS measurementsTable 5. Relative and by accuracy the digital of attitude measurements and length performed measurements on the model of DOM1 (Table and 5) DOM2..

Table 5. Relative accuracy of attitude and length measurements of DOM1 and DOM2. Number of Azimuth Difference Plunge Difference Length Difference DOM GCPs Couples Azimuth Number of Mean St. Dev.Plunge Mean Difference St. Dev. Length Mean Difference St. Dev. DOM ◦Difference ◦ ◦ ◦ 1GCPs 15 Couples 0.8 1.3 0.9 1.1 0.3% 0.1% 2 10Mean 0.4◦ St. Dev.0.5◦ Mean0.6◦ St. Dev.0.3 ◦ Mean0.2% St. Dev. 0.3% 1 15 0.8° 1.3° 0.9° 1.1° 0.3% 0.1% 2 10 0.4° 0.5° 0.6° 0.3° 0.2% 0.3% 2. the attitudes of 10 control planes (5 planes on each outcrop, see Figure6) accurately measured in 2.the the field attitudes and in of the 10 stereo-vision control planes of (5 DOMsplanes on (Table each6 outcrop). , see Figure 6) accurately measured in the field and in the stereo-vision of DOMs (Table 6).

Figure 6. (a) Texture d digital outcrop mode l (TDOM), (b) triangular mesh and (c) fie ld surve y picture Figure 6. (a) Textured (TDOM), (b) triangular mesh and (c) field survey picture of the same portion of the outcrop 1 (peoples for scale). In (a,c) the surface representing the control of theplane same ‘C’ portion are shaded of the in outcrop purple. 1In (peoples (b) the forbest scale).-fit plane In (calculatea,c) the surface d using representingCloud Compare the are control planerepresented ‘C’ are shaded by ain purple purple. rectangle. In (b)the best-fit plane calculated using Cloud Compare are represented by a purple rectangle.

Geosciences 2018, 8, 299 9 of 16

Table 6. Comparison between field manual attitude measures (O1 and O2) and DOM digital attitude Geosciences 2018, 8, x FOR PEER REVIEW 9 o f 16 measures (DOM1 and DOM2). The angular difference represents the solid angle between the attitude vectors. Table 6. Comparison betwe en fie ld manual attitude measures (O1 and O2) and DOM digital attitude measures (DOM1 and DOM2). The angular difference represents the solid angle between the attitude Angular Angular Surface Label O1 DOM1 (UAVDP) Surface Label O2 DOM2 (TDP) vectors. Difference Difference ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ A (fracture) 1 045 N 87 046 N 86 1 F (bed) 2 227 N 80 227 N 78 Angular2 Surface Label O1 DOM1 (UAVDP ) Angular Difference Surface Label O2 DOM2 (TDP ) B (fracture) 1 130◦ N 43◦ 129◦ N 45◦ 2◦ G (bed) 2 238◦ N 39◦ 237◦ N 40◦ Difference1◦ 1 ◦ ◦ ◦ ◦ ◦ 2 ◦ ◦ ◦ ◦ ◦ CA (bed)(fracture) 1 046045° N 87° 86 046°046 N N86° 84 1°2 HF (bed) (be d) 2 257227° N 80° 47 227°255 N N78° 46 2° 2 ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ DB (bed)(fracture1 ) 1 318130° N 43° 84 129°318 N N45° 82 2°2 IG (bed) (be d)2 2 239238° N 39° 26 237°239 N N40° 26 1° 0 ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ E (bed)C (be d)1 1 225046° N 86° 85 046°226 N N84° 85 2°1 JH (bed) (be d)2 2 234257° N 47° 22 255°235 N N46° 22 2° 1 D (be d) 1 318° N 84° 318° N 82° 2° I (be d) 2 239° N 26° 239° N 26° 0° 1 Control planes measured on the outcrop1; 2 Control planes measured on the outcrop2. E (be d) 1 225° N 85° 226° N 85° 1° J (bed) 2 234° N 22° 235° N 22° 1° 1 Control planes measured on the outcrop1; 2 Control planes measured on the outcrop2. The results indicate a high degree of relative accuracy of the model, with maximum angular differencesThe results in the attitudeindicate a and high length degree of of the relative vectors accuracy encompassed of the model, in ±2 with◦ and maximum 0.3%, respectively. angular Thedifferences comparison in the of attitude field and and DOMs length measures of the vectors (Table encompassed6) shows a in maximum ±2° and 0.3%, angular respectively difference. The of ~2◦comparisonfor both DOM1 of field and and DOM2. DOMs measures As previously (Table outlined, 6) shows thesea maximum values angular cannot bediff considerederence of ~2° errors for becauseb ot h DOM manual1 and sampling DOM2. couldAs previously be affected outlined, by orientation these values bias cannot due to be the considered local variation errors ofbecause surface orientations,manual sampling whereas could DOM be sampling affected could by orientation overcome orientationbias due to bias the collecting local variation the orientation of surface that bestorientations, fits the entire whereas surface. DOM These sampling angular could differences overcome confirm orientation the validity bias collecting of the DOMs the orientation and their that good relativebest fits accuracy. the entire On surface. the whole, These theangular substantial differences agreement confirm betweenthe validity field of the manual DOMs and and DOM their good digital relative accuracy. On the whole, the substantial agreement between field manual and DOM digital measures largely counterbalances the not completely satisfying absolute accuracy of the models; the measures largely counterbalances the not completely satisfying absolute accuracy of the models; the latter is probably due to the few measurable GCPs that are concentrated essentially at the base of the latter is probably due to the few measurable GCPs that are concentrated essentially at the base of the rock slopes. rock slopes. 5. Results 5. Results 5.1. Folds Analysis 5.1. Folds Analysis A series of NE-vergent overturned folds were identified by means of stereoscopic analysis and A series of NE-vergent overturned folds were identified by means of stereoscopic analysis and interpretation of textured DOMs. The DOMs were subdivided into domains corresponding to different interpretation of textured DOMs. The DOMs were subdivided into domains corresponding to limbs.DOM1 was subdivided into 4 main fold limb domains (Figure7), whereas DOM2 was subdivided different limbs.DOM1 was subdivided into 4 main fold limb domains (Figure 7), whereas DOM2 was into 2 main fold limb domains (Figure8). subdivided into 2 main fold limb domains (Figure 8).

FigureFigure 7. 7Structural. Structural interpretation interpretation of DOM DOM 1 1(see (see Figure Figure 4a)4a) projected projected on a onn orthorectifie an orthorectified d image image.. The Thevie view w point point of of the the orthore orthorectified ctifie d image image is is horizontal horizontal and and sub sub-orthogonal-orthogonal to tothe the mean mean slope slope dir direction.e ct ion. BedsBe ds(thin (thin black black lines) line s) and and thethe main traces traces of of the the fold fold axial axial plane planes s (grey (grey dashed dashed line s) lines) are shown. are shown. The Thenumbe numbers rs mark mark the the 4 4 main main fold fold limbs limbs re recognized cognize d by by the the pre preliminary liminary visual visual analysis. analysis. Small Small black black arrowsarrows show show the the polarity polarity of of the the beds. beds.

Geosciences 2018, 8, 299 10 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 10 o f 16

FigureFigure 8.8. StructuralStructural interpretation interpretation of of DOM DOM 2 (see 2 (see Figure Figure 4b)4b) projected projected on on a orthorectifie a orthorectified d im a image. ge . The Thevie w view point point of the of orthore the orthorectified ctifie d image image is horizontal is horizontal and sub and-orthogonal sub-orthogonal to the mean to the slope mean direction. slope direction.Be ds (thin Beds black (thin line black s) and lines) the andmain the trace main of trace the fold of the axial fold pl axialane plane(grey (greydashed dashed line) line)are shown. are shown. The Thenumbe numbers rs mark mark the the2 main 2 main fold fold limbs limbs re recognized cognize d by by the the pre preliminary liminary visual visual analysis. Small blackblack arrowsarrows showshow thethe polaritypolarity ofof thethe beds.beds.

InIn DOM1DOM1 the the limb limb 4 (L4;4 (L4; Figure Figure5) is 5) formed is formed by sub-parallel by sub-parallel beds dipping beds dipping towards towards SW. The SW. surface The morphology,surface morphology, colors and colors shape and of theshape beds of suggestthe beds a normalsuggest polaritya normal of thepolarity strata. of Thisthe strata. hypothesis This washypothesis confirmed was by confirmed field controls. by field By controls. the analysis By the of analysis DOM1 29 of bedsDOM1 of 29 L4 beds were of measured L4 were measured and their orientationand their orientation statistics were statistics calculated were calculated (see Table7 (see). Table 7).

TableTable 7. 7. FoldFold limb limb me mean an attitude attitude (dip dire direction ction and dip) and K-FisherK-Fishe r dispersiondispe rsion coefficient. coe fficie nt. Fo ld Limb Denomination L0 L1 1 L2 L3 L4 Fold LimbN° Denomination o f Measured Beds L0 15 L1201 1 22L2 7 L329 L4 N◦ of MeasuredMean Attitude Beds 15235° N 24° 228°20 1N 80° 246° N22 22° 037° N 76° 7 244° N 36° 29 Mean AttitudeK-Fisher Co effic ient 235◦ N11.79 24◦ 228◦19.72N 80 ◦ 24629.65◦ N 22◦ 82.36037 ◦ N 7627.61◦ 244 ◦ N 36◦ K-Fisher Coefficient 11.791 Overturned 19.72 beds. 29.65 82.36 27.61 1 Overturned beds. The bedding has a mean attitude (dip direction and dip) of 244° N 36°. L4 and L3 form a NE- vergentThe beddingoverturned has asyncline mean attitude (Figure (dip 7). direction This fold and (F3 dip)-4) of decreases 244◦ N 36◦ .of L4 magnitude and L3 form upward a NE-vergent until overturneddisappears. syncline (Figure7). This fold (F3-4) decreases of magnitude upward until disappears. OnlyOnly 77 bedbed orientations orientations of of L3 L3 were were sampled sampled for for its limitedits limited dimensions. dimensions. The meanThe mean attitude attitude of these of bedsthese is beds 37◦ Nis 7637°◦ (seeN 76° Table (see7 ).Table L3 and 7). theL3 lowerand the part lower of L2 part form of anL2 overturned form an overturned NE-vergent NE anticline-vergent (F2-3;anticline Figure (F27-3).; Figure 7). AA totaltotal ofof 2222 bedbed orientationsorientations ofof L2L2 werewere sampled.sampled. TheThe meanmean attitudeattitude ofof L2L2 bedsbeds isis 246246°◦ NN 2222°◦ (see(see TableTable7 ).7). L2 L2 and and L1 L1 form form a a NE-vergent NE-vergent overturned overturned syncline syncline (F1-2; (F1-2 Figure; Figure7). 7). L1L1 is characterized by overturned beds beds with with a a mean mean attitude attitude of of 228° 228 N◦ N 80°. 80 ◦ Limb. Limb L1 L1is visible is visible on onDOM1 DOM1 and and DOM2 DOM2 (see (see Figure Figures 7 7and and 88);); the the analysis analysis performed onon both both the the models models allowed allowed the the acquisitionacquisition ofof thethe measuresmeasures ofof 2020 bedsbeds (Table(Table7 ).7). AA total total ofof 1515 bedbed orientationsorientations werewere measuredmeasured alsoalso onon LimbLimb L0L0 thatthat formsforms withwith limblimb L1L1 anotheranother NE-vergentNE-vergent overturnedoverturned anticlineanticline (F0-1).(F0-1). HereHere thethe meanmean attitudeattitude of of thethe bedsbeds isis 235235°◦ N N 2424°◦ (Table(Table7 7).). TheThe digitaldigital measurementsmeasurements ofof beddingbedding attitude of each foldfold limbslimbs (Figure(Figure9 9)) and and their their statistic statistic parametersparameters suchsuch as meanmean dipdip direction,direction, dip and K-FisherK-Fisher distribution coefficientcoefficient (see Table7 ),7), allowedallowed toto calculatecalculate thethe orientationorientation ofof foldfold axes axes andand axialaxial planesplanes andand thethe differentdifferent interlimbinterlimb anglesangles (Figure(Figure 1010;; TableTable8 ).8).

Geosciences 2018, 8, 299 11 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 11 o f 16 Geosciences 2018, 8, x FOR PEER REVIEW 11 o f 16

FigureFigure 9. 9. LowerLowe r he hemisphere misphe re stereographic stereographic projections projections of ofbedding bedding plane planes s and andpole poless of limb of 0 limb (a), limb 0 (a ), Figure 9. Lowe r he misphe re stereographic projections of bedding plane s and pole s of limb 0 (a), limb limb1 (b 1), (limbb), limb 2 (c), 2 limb (c), limb 3 (d) 3 and (d) andlimb limb 4 (e). 4 (e). 1 (b), limb 2 (c), limb 3 (d) and limb 4 (e).

Figure 10. Lowe r he misphe re stereographic projections of fold axial plane s (AP) and fold axe s (A). FigureFigure 10. 10. Lowe r he misphe re stereographic projections of fold axial plane s (AP) and fold axe s (A). Folds are Lowerindicated hemisphere by the connected stereographic limbs (e.g., projections Fold 0-1 means of fold fold axial forms planes by limbs (AP) and0 and fold 1). Attitude axes (A). FoldsFolds are are indicated indicated by by the the connected connected limbs limbs (e.g., (e.g., Fold Fold 0-10-1 means fold forms forms by by limbs limbs 0 0and and 1). 1). Attitude Attitude datadata are are shown shown inin TableTable 8. data are shown in Table8.

TableTable 8.8. CalculateCalculate d d attitudes of Axe s (A) and AxialAxial Plane Plane s s (AP) (AP) and and inte inte rlimb rlimb angle angle s sof of the the e e xamined xamined Table 8. Calculated attitudes of Axes (A) and Axial Planes (AP) and interlimb angles of the foldsfolds.. examined folds. FoFo ld Denomination F0F0--11 F1F1-2-2 F2F2-3-3 F3F3-4-4 APAP AttitudeAttitude (Dip Direction and Dip)) 230°230° N N 47° 47° 233°233° N N 34° 34° 230°230° N N 49° 49° 239°239° N N 44° 44° Fold Denomination F0-1 F1-2 F2-3 F3-4 A Azimuth [°] 313313 328328 312312 317317 AP Attitude (Dip A Plung e 230[°] ◦ N 47◦ 23366 ◦ N 34◦ 00 230◦ N9 499 ◦ 2391212 ◦ N 44◦ Direction andInterlimbInterlimb Dip) Angle [°] 5656 5959 9696 108108 A Azimuth [◦] 313 328 312 317 A Plunge [◦] 6 0 9 12 TheThe folds folds show show veryvery similar features; they areare endowedendowed with with sub sub--horizontalhorizontal or or slightly slightly NW NW Interlimb Angle [◦] 56 59 96 108 dippingdipping axesaxes andand SWSW dipping axial planes, indicatingindicating thatthat they they belongbelong to to the the same same deformation deformation phasephase. . TheTheThe folds geometry geometry show ofof very the the investigated similar features; asymmetric they are folds folds endowed (z (z--shapedshaped with looking looking sub-horizontal towards towards NW) NW) or indicates slightly indicates NWa a dippingvergencevergence axes towardstowards and SW NENE dipping andand confirms axial planes, the tectonic indicating transporttransport that theytowards towards belong NE NE toof of the the sameAntola Antola deformation Unit, Unit, driven driven phase. by by thetheThe Levanto Levanto geometry--OttoneOttone of thrust,thrust, the investigated which is located asymmetric immediately folds NE NE (z-shaped of of the the area area looking (Figures (Figures towards 1 1and and 2). NW) 2). indicates a vergence towards NE and confirms the tectonic transport towards NE of the Antola Unit, driven by the5.2.5.2. Levanto-Ottone Fracture Fracture AAnalysisnalysis thrust, which is located immediately NE of the area (Figures1 and2). FractureFracture analysisanalysis was conducted on each limblimb ofof thethe foldsfolds visiblevisible on on the the two two DOMs. DOMs. 213, 213, 119, 119, 5.2. Fracture Analysis 138,138, 30, 30, 184 184 fracturesfractures werewere measured on limbs L0L0 toto L4,L4, respectively.respectively. The The identified identified fractures fractures can can be be subdividedsubdividedFracture intointo analysis 4 4 setssets was (Tables(Tables conducted 9–12). on each limb of the folds visible on the two DOMs. 213, 119, 138, 30,FractureFracture 184 fractures analysis analysis were shows measured that all onthe limbs sets ofof L0 fracture fracture to L4, are respectively.are generally generally strata Thestrata identified bound bound and and fractures always always sub can sub- be- subdividedorthogonalorthogonal into to to the the 4 sets bedding;bedding; (Tables besides9–12). the angular relationshipsrelationships among among these these four four sets sets and and those those between between thisthisFracture network network analysisof of fractures fractures shows andand the that bedding all the planesplanes sets areare of fracturealmost almost the the are same same generally for for all all the the strata fold fold limbs limbs bound (Figure (Figure and 11). always11). sub-orthogonal to the bedding; besides the angular relationships among these four sets and those between this network of fractures and the bedding planes are almost the same for all the fold limbs (Figure 11).

Geosciences 2018, 8, 299 12 of 16

Table 9. Features of fracture set K2 determined in the different fold limbs.

Fold Limb Mean Attitude (Dip Number of Fractures K-Fisher Coefficient Denomination Direction and Dip) L0 13 133◦ N 80◦ 43.22 L1 12 123◦ N 73◦ 41.21 L2 10 182◦ N 86◦ 41.95 L3 1 306◦ N 86◦ - L4 32 147◦ N 85◦ 45.93

Table 10. Features of fracture set K3 determined in the different fold limbs.

Fold Limb Mean Attitude(Dip Number of Fractures K-Fisher Coefficient Denomination Direction and Dip) L0 27 039◦ N 73◦ 31.04 L1 16 022◦ N 16◦ 94.70 L2 31 077◦ N 71◦ 43.67 L3 12 193◦ N 22◦ 25.91 L4 23 066◦ N 66◦ 58.76

Table 11. Features of fracture set K4 determined in the different fold limbs.

Fold Limb Mean Attitude(Dip Number of Fractures K-Fisher Coefficient Denomination Direction and Dip) L0 55 349◦ N 85◦ 308.7 L1 20 103◦ N 35◦ 52.58 L2 34 047◦ N 82◦ 68.76 L3 2 292◦ N 50◦ - L4 24 002◦ N 72◦ 42.57

Table 12. Features of fracture set K5 determined in the different fold limbs.

Fold Limb Mean Attitude(Dip Number of Fractures K-Fisher Coefficient Denomination Direction and Dip) L0 46 095◦ N 68◦ 37.54 L1 19 313◦ N 71◦ 39.15 L2 50 134◦ N 74◦ 51.85 L3 13 136◦ N 66◦ 47.59 L4 42 113◦ N 70◦ 37.44

These features allow for correlating the fracture sets of the different fold limbs considering that sets K2 and K3 are sub-orthogonal and sub-parallel to the fold axes and to the directions of the bedding planes, while sets K4 and K5 are oblique. The set K3 can be interpreted as a joint set striking parallel to the fold hinge and dipping normal to bedding. It may be associated with outer-arc stretching during folding; regions of localized tensional stresses develop in the same orientation as regional compression, leading to fracture opening. The K2 joint set, which strikes perpendicular to the fold hinge and dips normal to bedding may be associated with localized extension in a direction perpendicular to the tectonic transport. The remaining fractures associated with these thrust related folds are two sets of conjugate shear fractures (K4 and K5) generally with an acute bisector parallel to the tectonic transport direction. These fractures may form due to regional compression or localized inner-arc compression associated with tangential longitudinal strain folding (Figure 12). Incomplete assemblages of these fractures are seen at limb3 where only one K2-fracture is visible. Geosciences 2018, 8, 299 13 of 16 Geosciences 2018, 8, x FOR PEER REVIEW 13 o f 16

Geosciences 2018, 8, x FOR PEER REVIEW 13 o f 16

Figure 11. Contour plot of poles (equal area, lower hemisphere, per 1% area) and cyclographic Figure 11. Contour plot of poles (equal area, lower hemisphere, per 1% area) and cyclographic projection of the mean planes (equal angle, lower hemisphere) of bedding and fractures of limbs 0 (a), Figureprojection 11. ofContour the mean plot planes of poles (equal (equal angle, area, lower lower hemisphere) hemisphere, of bedding per 1% and area) fractures and cyclographic of limbs 0 ( a), 1 (b), 2 (c), 3 (d) and 4 (e). projection1 (b), 2 (c), of 3 (thed) andmean 4 (plane). es (equal angle, lower hemisphere) of bedding and fractures of limbs 0 (a), 1 (b), 2 (c), 3 (d) and 4 (e).

Figure 12. A 3D interpretation of the geometrical relationships between fracture sets and folds (not to Figurescale ). 12. A 3D interpretation of the geometrical relationships between fracture sets and folds (not to Figure 12. A 3D interpretation of the geometrical relationships between fracture sets and folds scale ). 6. Concluding(not to scale). Remarks 6. Concluding Remarks Our study demonstrates that the Digital Photogrammetry (DP) techniques and in particular thoseOur based study on Unmanneddemonstrates Aerial that Vehiclethe Digital (UAV) Photogrammetry technologies can (DP) produce techniques Digital and Outcrop in particular Models those based on Unmanned Aerial Vehicle (UAV) technologies can produce Digital Outcrop Models

Geosciences 2018, 8, 299 14 of 16

6. Concluding Remarks Our study demonstrates that the Digital Photogrammetry (DP) techniques and in particular those based on Unmanned Aerial Vehicle (UAV) technologies can produce Digital Outcrop Models (DOMs) of high accuracy and low costs that can be successfully applied to detailed structural analysis of rock outcrops. The principal advantages of this technique are: (1) the possibility to analyze outcrops, totally or partially inaccessible; (2) obtaining a largest number of measures and data; (3) acquiring more representative measures of geological planes (strata, faults and fractures) with respect to manual sampling, because the measurements are not influenced by the local variations of attitude that typically affect the geological planes; (4) the possibility to repeat and control the analysis and the measures by different operators and at different times; (5) a substantial time savings in the phase of the fracture measurements; (6) a safe methodology especially in case of unstable and dangerous rock slope. During this study two large inaccessible outcrops of the Monte Antola Formation, a turbiditic calcareous flysch, located in NW Apennines, near Ponte Organasco village along the Trebbia river, have been studied. From DOMs analysis, it has been possible to measure the attitude of the bedding of the whole outcrops and then calculate with good precision the orientation of fold axes and axial planes. The folds analyzed were scarcely visible and not measurable by field surveys due to the steepness (more than 80◦) and height (several tens of meters) of the outcrops. The folds are constituted by two couples of NE-vergent overturned anticlines and synclines. The folds are asymmetric (z-shaped looking towards NW), rather irregular and show major variations of thickness and shape in the hinge-regions. The interlimb angles vary from few degrees (fold between limbs 0 and 1) to about 90◦ (fold between limbs 3 and 2). The axes and axial planes of the folds have a similar orientation: the mean attitude of the fold axes is 319◦ N 5◦, while the mean orientation of the axial planes is 233◦ N 42◦. The fold analysis indicates that the deformations affecting the Monte Antola Formation in this area have a typical Apenninic trend, with a clear NE vergence, confirming the presence and nature of the Levanto-Ottone thrust. Similarly, the analysis of the two DOMs allowed to measure the fracture network that locally affects the flysch and to evaluate the relationships between fractures and folds, examining the whole development of the structures, from the limbs to the hinges. Also, these measurements were not achievable by field surveys for the inaccessibility of the outcrops. All the limbs of the folds are affected by four main sets of fractures (K2, K3, K4 and K5). These sets are always sub-orthogonal to the bedding and apparently maintain the same angular relationships both among them and with the bedding in the different parts of the folds, even where the strata are sub-vertical or overturned (limb 1). This condition suggests that fractures formed in the early stages of the deformation, before the complete enucleation of the folds. In particular, considering the fractures genetically connected to the folding, set K2, that is sub-orthogonal to the fold planes, can be considered a cross/extensional/hinge-orthogonal joint set, set K3, that has a direction sub-parallel to the fold axes, can be considered a longitudinal/radial joint/hinge-parallel set, while sets K4 and K5 can be interpreted as conjugate shear fractures oblique with respect to the fold axial planes and axes, generally with an acute bisector often parallel to the thrust transport direction (Figure 12).

Author Contributions: Performed the UAV and RTK GPS surveys and elaborate raw data, M.C. Conceptualized the study, Developed the methodology, Performed formal analysis, Prepared the initial draft, N.M. and C.P. Supervised and provided critical reviews and editing, C.M. and C.P. Funding: This research received no external funding. Acknowledgments: Many thanks are due to Ecates S.r.l. for the support on the fly of the UAV, to Daniel Girardeau-Montaut and other CloudCompare developers that developed the useful open-source software and continue to implement it. Conflicts of Interest: The authors declare no conflict of interest. Geosciences 2018, 8, 299 15 of 16

References

1. Powers, P.S.; Chiarle, M.; Savage, W.Z. A digital photogrammetric method for measuring horizontal surficial movements on the Slumgullion earthflow, Hinsdale County, Colorado. Comput. Geosci. 1996, 22, 651–663. [CrossRef] 2. Xu, X.; Aiken, C.L.; Bhattacharya, J.P.; Corbeanu, R.M.; Nielsen, K.C.; McMechan, G.A.; Abdelsalam, M.G. Creating virtual 3-D outcrop. TLE Soc. Explor. Geophys. 2000, 19, 197–202. [CrossRef] 3. Pringle, J.K.; Westerman, A.R.; Clark, J.D.; Drinkwater, N.J.; Gardiner, A.R. 3D high-resolution digital models of outcrop analogue study sites to constrain reservoir model uncertainty: An example from Alport Castles, Derbyshire, UK. Petrol. Geosci. 2004, 10, 343–352. [CrossRef] 4. Bellian, J.A.; Kerans, C.; Jennette, D.C. Digital outcrop models: Applications of terrestrial scanning lidar technology in stratigraphic modeling. J. Sediment. Res. 2005, 75, 166–176. [CrossRef] 5. Sturzenegger, M.; Stead, D. Close-range terrestrial digital photogrammetry and terrestrial laser scanning for discontinuity characterization on rock cuts. Eng. Geol. 2009, 106, 163–182. [CrossRef] 6. Jaboyedoff, M.; Oppikofer, T.; Abellán, A.; Derron, M.H.; Loye, A.; Metzger, R.; Pedrazzini, A. Use of LIDAR in landslide investigations: A review. Nat. Hazards 2012, 61, 5–28. [CrossRef] 7. Westoby, M.J.; Brasington, J.; Glasser, N.F.; Hambrey, M.J.; Reynolds, J.M. ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 2012, 179, 300–314. [CrossRef] 8. Humair, F.; Pedrazzini, A.; Epard, J.-L.; Froese, C.R.; Jaboyedoff, M. Structural characterization of turtle mountain anticline (Alberta, Canada) and impact on rock slope failure. Tectonophysics 2013, 605, 133–148. [CrossRef] 9. Bemis, S.P.; Micklethwaite, S.; Turner, D.; James, M.R.; Akciz, S.; Thiele, S.T.; Bangash, H.A. Ground-based and UAV-based photogrammetry: A multi-scale, high-resolution mapping tool for structural geology and paleoseismology. J. Struct. Geol. 2014, 69, 163–178. [CrossRef] 10. Spreafico, M.C.; Francioni, M.; Cervi, F.; Stead, D.; Bitelli, G.; Ghirotti, M.; Girelli, V.A.; Lucente, C.C.; Tini, M.A.; Borgatti, L. Back Analysis of the 2014 San Leo Landslide Using Combined Terrestrial Laser Scanning and 3D Distinct Element Modelling. Rock Mech. Rock Eng. 2016, 49, 2235–2251. [CrossRef] 11. Tavani, S.; Corradetti, A.; Billi, A. High precision analysis of an embryonic extensional fault-related fold using 3D orthorectified virtual outcrops: The viewpoint importance in structural geology. J. Struct. Geol. 2016, 86, 200–210. [CrossRef] 12. Remondino, F.; El-Hakim, S. Image-based 3D modelling: A review. Photogramm. Rec. 2006, 21, 269–291. [CrossRef] 13. Colomina, I.; Molina, P. Unmanned aerial systems for photogrammetry and remote sensing: A review. ISPRS J. Photogramm. Remote Sens. 2014, 92, 79–97. [CrossRef] 14. Niethammer, U.; James, M.R.; Rothmund, S.; Travelletti, J.; Joswig, M. UAV-based remote sensing of the Super-Sauze landslide: Evaluation and results. Eng. Geol. 2012, 128, 2–11. [CrossRef] 15. Lucieer, A.; Jong, S.M.D.; Turner, D. Mapping landslide displacements using Structure from Motion (SfM) and image correlation of multi-temporal UAV photography. Prog. Phys. Geogr. 2014, 38, 97–116. [CrossRef] 16. Tannant, D.D. Review of Photogrammetry-Based Techniques for Characterization and Hazard Assessment of Rock Faces. Int. J. Geohazards Environ. 2015, 1, 76–87. [CrossRef] 17. Casella, E.; Rovere, A.; Pedroncini, A.; Stark, C.P.; Casella, M.; Ferrari, M.; Firpo, M. Drones as tools for monitoring beach topography changes in the Ligurian Sea (NW Mediterranean). Geo-Mar. Lett. 2016, 36, 151–163. [CrossRef] 18. Salvini, R.; Mastrorocco, G.; Seddaiu, M.; Rossi, D.; Vanneschi, C. The use of an unmanned aerial vehicle for fracture mapping within a marble quarry (Carrara, Italy): Photogrammetry and discrete fracture network modelling. Geomat. Nat. Hazards Risk 2016, 8, 34–52. [CrossRef] 19. Sturzenegger, M.; Stead, D.; Elmo, D. Terrestrial remote sensing-based estimation of mean trace length, trace intensity and block size/shape. Eng. Geol. 2011, 119, 96–111. [CrossRef] 20. Levi, N.; Ellero, A.; Ottria, G.; Pandolfi, L. Polyorogenic deformation history recognized at very shallow structural levels: The case of the Antola Unit (Northern Apennine, Italy). J. Struct. Geol. 2006, 28, 1694–1709. [CrossRef] Geosciences 2018, 8, 299 16 of 16

21. Catanzariti, R.; Ellero, A.; Levi, N.; Ottria, G.; Pandolfi, L. Calcareous nannofossil biostratigraphy of the Antola Unit succession (Northern Apennines, Italy): New age constraints for the Late Cretaceous Helminthoid Flysch. Cretaceous Res. 2007, 28, 841–860. [CrossRef] 22. Marroni, M.; Meneghini, F.; Pandolfi, L. Anatomy of the Ligure-Piemontese subduction system: Evidence from Late Cretaceous–middle Eocene convergent margin deposits in the Northern Apennines, Italy. Int. Geol. Rev. 2010, 52, 1160–1192. [CrossRef] 23. Cavanna, F.; Di Giulio, A.; Galbiati, B.; Mosna, S.; Perotti, C.R.; Pieri, M. Carta geologica dell’estremità orientale del Bacino Terziario Ligure-Piemontese. Atti Ticinensi di Scienze della Terra 1989, 32, 191–195. (In Italian) 24. Elter, P. L’ensemble Ligure. Available online: http://www.beactiveliguria.it/fr/beactive-fr/%C3%A0-pieds. html (accessed on 8 August 2018). 25. Marroni, M.; Molli, G.; Montanini, A.; Ottria, G.; Pandolfi, L.; Tribuzio, R. The External Liguride units (Northern Apennine, Italy): From rifting to convergence history of a fossil ocean–continent transition zone. Ofioliti 2002, 27, 119–132. 26. Birch, J.S. Using 3DM Analyst mine mapping suite for rock face characterization. In Proceedings of the 41st U.S. Rock Mechanics Symposyum: Laser and Photogrammetric Methods for Rock Face Characterization, Asheville, NC, USA, 28 June–1 July 2009; pp. 13–32. 27. Spetsakis, M.; Aloimonos, J.Y. A multi-frame approach to visual motion perception. Int. J. Comput. Vis. 1991, 6, 245–255. [CrossRef] 28. Boufama, B.; Mohr, R.; Veillon, F. Euclidean constraints for uncalibrated reconstruction. In Proceedings of the Fourth International Conference on Computer Vision, Berlin, German, 11–14 May 1993; pp. 466–470. 29. Agisoft Photoscan (Version 1.2.5). Available online: http://www.agisoft.com/ (accessed on 7 August 2018). 30. CloudCompare (Version 2.9). Available online: http://www.danielgm.net/cc/ (accessed on 7 August 2018). 31. Thiele, S.T.; Grose, L.; Samsu, A.; Micklethwaite, S.; Vollgger, S.A.; Cruden, A.R. Rapid, semi-automatic fracture and contact mapping for point clouds, images and geophysical data. Solid Earth 2017, 8, 1241–1253. [CrossRef] 32. Chesley, J.T.; Leier, A.L.; White, S.; Torres, R. Using Unmanned Aerial Vehicles and Structure-From-Motion Photogrammetry to Characterize Sedimentary Outcrops: An Example from the Morrison Formation, Utah, USA. Sediment. Geol. 2017, 354, 1–8. [CrossRef] 33. Cawood, A.J.; Bond, C.E.; Howell, J.A.; Butler, R.W.; Totake, Y. LiDAR, UAV or compass-clinometer? Accuracy, coverage and the effects on structural models. J. Struct. Geol. 2017, 98, 67–82. [CrossRef] 34. Jordá Bordehore, L.; Riquelme, A.; Cano, M.; Tomás, R. Comparing manual and remote sensing field discontinuity collection used in kinematic stability assessment of failed rock slopes. Int. J. Rock Mech. Min. Sci. 2017, 97, 24–32. [CrossRef]

© 2018 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/).