applied

sciences

ArticleArticle KnudKnud RasmussenRasmussen GlacierGlacier StatusStatus AnalysisAnalysis BasedBased onon HistoricalHistorical DataData andand MovingMoving DetectionDetection UsingUsing RPASRPAS

KarelKarel PavelkaPavelka **, Jaroslav, Jaroslav Šedina Šedina and and Karel Karel Pavelka Pavelka,, Jr Jr..

FacultyFaculty ofof CivilCivil Engineering,Engineering, CzechCzech TechnicalTechnical University University in in Prague, Prague, Thakurova Thakurova 7, 7, 16629 16629 Prague, Prague Czech, Republic; [email protected] Republic; [email protected] (J.Š.); [email protected] (J.Š.) [email protected] (K.P.J.) (K.P.J.) ** Correspondence:Correspondence: [email protected]; [email protected]; Tel.: Tel.: +420-608-211-360 +420-608-211-360

Abstract:Abstract: ThisThis articlearticle discussesdiscusses partialpartial resultsresults ofof anan internationalinternational scientificscientific expeditionexpedition toto GreenlandGreenland thatthat researchedresearched thethe geography,geography, geodesy,geodesy, botany,botany, andand glaciologyglaciology ofof thethe area.area. TheThe resultsresults herehere focusfocus onon the photogrammetrical results results obtained obtained with with the the eBee eBee drone drone in in the the eastern eastern part part ofof Greenland Greenland at atthe the front front of ofthe the Knud Knud Rasmussen Rasmussen Glacier Glacier and and the theuse useof archive of archive image image data data for monitoring for monitoring the con- the conditiondition of this of thisglacier glacier.. In these In theseshort- short-termterm visits to visits the site, to the the site, possibility the possibility of using ofa drone using is a discussed drone is discussedand the results and the show results not showonly the not flow only thespeed flow of speedthe glacier of the but glacier alsobut the alsoshape the and shape structure and structure from a fromheight a heightof up to of 200 up m. to From 200 m. two From overflights two overflights near the near glacier the front glacier at frontdifferent at different times, it times,was possible it was possibleto obtain to the obtain speed the of speed the glacier of the glacierflow and flow the and distribution the distribution of velocities of velocities in the in glacier the glacier stream. stream. The Thetechnology technology uses usesa comparison a comparison of two of point two pointclouds clouds derived derived from a fromset of aaerial set of photos aerial taken photos with taken the witheBee thedrone, eBee and drone, calculating and calculating the M3C2 the (M M3C2ultiscale (Multiscale Model-to- Model-to-ModelModel Cloud Comparison Cloud Comparison)) distances distanceswith CloudCompare with CloudCompare software. software.The results The correlate results correlatewith other with measurement other measurement methods methodslike accurate like accurateand long and-term long-term measurement measurement with Global with Global Navigation Navigation Satellite Satellite System System (GNSS (GNSS),), satellite satellite radar, radar, or orground ground geodetical geodetical technology. technology. The The resulting resulting speed speed from from the thedrone drone data data reached reached in the in middle the middle part partof the of glacier the glacier,, was wasapproximately approximately 12–15 12–15mm per day. per day.The second The second part partof the of paper the paper focuses focuses on the on anal- the analysisysis of modern of modern satellite satellite images images of the of Knud the Knud Rasmussen Rasmussen Glacier Glacier from fromGoogle Google Earth Earth(Landsat (Landsat series


series1984–2016) 1984–2016) and Sentinel and Sentinel 2a, and 2a, a andcomparison a comparison with historical with historical aerial aerial images images from from1932 1932to 1933. to 1933. His- Citation:
 Pavelka, K.; Šedina, J.; Historicaltorical images images were were processed processed photogrammetrically photogrammetrically into into a athree three-dimensional-dimensional ( (3D)3D) model. Finally, Pavelka, K. Jr. Knud Rasmussen Citation: Pavelka, K.; Šedina, J.; orthogonalizedorthogonalized imageimage datadata fromfrom threethree systemssystems (drone(drone photos,photos, historicalhistorical aerialaerial photos,photos, andand satellitesatellite Glacier Status Analysis Based on Pavelka, K., Jr. Knud Rasmussen data)data) werewere comparedcompared inin thethe ArcGISArcGIS software.software. ThisThis allowsallows usus toto analyzeanalyze glacierglacier changeschanges overover timetime Historical Data and Moving Glacier Status Analysis Based on inin thethe timetime spanspan fromfrom 19321932 toto 2020,2020, withwith thethe caveatcaveat thatthat fromfrom 19331933 toto 19831983 wewe diddid notnot havehave datadata atat Detection Using RPAS. Appl. Sci. Historical Data and Moving Detection 2021, 11, x. ourour disposal.disposal. TheThe resultresult showsshows thatthat moremore significantsignificant changeschanges inin thethe areaarea ofof thisthis glacierglacier occurredoccurred afterafter Using RPAS. Appl. Sci. 2021, 11, 754. https://doi.org/10.3390/xxxxx 2011.2011. TheThe mainmain aimaim ofof thisthis articlearticle isis to research thethe useuse ofof photogrammetricphotogrammetric methodsmethods forfor monitoringmonitoring https://doi.org/10.3390/app11020754 thethe conditioncondition and and parameters parameters of of glaciers glaciers based based on on non-traditional non-traditional technology, technology such, such as drones as drones or new or Received: 30 November 2020 processingnew processing of historical of historical photos. photos. Received: 30 November 2020 Accepted: 12 January 2021 Accepted: 12 January 2021 Published: 14 January 2020 Keywords:Keywords: photogrammetry;photogrammetry; RPAS;RPAS; CloudCompare;CloudCompare; Bee;Bee; Greenland;Greenland; Knud Knud Rasmussen Rasmussen Glacier Glacier Published: 14 January 2021

Publisher’s Note: MDPI stays Publisher’s Note: MDPI stays neu- neutral with regard to jurisdictional tral with regard to jurisdictional clai- claims in published maps and 1. Introduction ms in published maps and institutio- 1. Introduction institutional affiliations. nal affiliations. II wouldwould likelike toto dedicatededicate thisthis articlearticle inin memoriam toto my friend,friend, scientist, and real man, ProfessorProfessor WilfriedWilfried KorthKorth (Figure(Figure1 ),1) who, who died died tragically tragically in in spring spring 2019, 2019, just just before before his his last last plannedplanned expeditionexpedition toto GreenlandGreenland(KP). (KP). Copyright: © 2021 by the authors. Copyright: © 2021 by the authors. Li- Submitted for possible open access censee MDPI, Basel, Switzerland. publication under the terms and This article is an open access article conditions of the Creative Commons distributed under the terms and con- Attribution (CC BY) license ditions of the Creative Commons At- (http://creativecommons.org/licenses tribution (CC BY) license (https:// /by/4.0/). creativecommons.org/licenses/by/ FigureFigure 1.1. ProfessorProfessor WilfriedWilfried KorthKorth (†(† 2019).2019). 4.0/).

Appl. Sci. 2021, 11, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/applsci Appl. Sci. 2021, 11, 754. https://doi.org/10.3390/app11020754 https://www.mdpi.com/journal/applsci Appl. Sci. 2021, 11, x FOR PEER REVIEW 2 of 19

The gradual melting of glaciers has been monitored for a long time. Since the 1980s, ice has been declining more than it is replenished in winter. Winters are milder and sum- Appl. Sci. 2021, 11, 754 mers longer and warmer. The ice that disappears will not be restored. Nowadays,2 of 19six times more ice has been disappearing from Greenland than in the 1980s [1]. The Greenland cap is a vast mass of ice covering 1.7 million square km, which repre- sentsThe about gradual 80% meltingof Greenland of glaciers’s surface. has been It is monitoredthe second forlargest a long glaciated time. Since area the in the 1980s, world ice ; hasthe beenfirst decliningis the Antarctic morethan cap. itIts is thickness replenished is usually in winter. more Winters than are2 km milder and andsometimes summers ex- longerceeds and3 km warmer. [2]. The The weight ice that of disappearsthe glacier willhas com not bepressed restored. the Nowadays,central part six of timesGreenland, more icebringing has been the disappearing rocky bedrock from below Greenland it to about than insea the level, 1980s while [1]. the mountain range sur- roundsThe the Greenland glacier almost cap is a along vastmass its entire of ice edge. covering This 1.7is detectable million square by the km, deformation which repre- of sentsthe Earth about’s 80%gravity of Greenland’s field. If the entire surface. Greenland It is the secondcap melted, largest the glaciated level of the area world in the’s world;oceans thewould first isrise the by Antarctic about 7 m cap.. Due Its to thickness the long is-term usually melting more of than the 2 glacier, km and the sometimes compressed exceeds rock 3is km gradually [2]. The weight rising onof the the glacier outskirts has compressed of Greenland. the According central part to of scientific Greenland, studies, bringing the theGreenland rocky bedrock coast rises below by 2 it.5 to cm about per year sea level,[3–5]. whileHowever, the mountainit is also scientifically range surrounds confirmed the glacierthat some almost parts along of the its glacier entire are edge. even This increasing. is detectable This byinformation the deformation indicates of that the Earth’sthe con- gravitydition of field. the IfGreenland the entire cap Greenland must continue cap melted, to be the carefully level of studied. the world’s Today, oceans scientific would satel- rise bylites about in particular 7 m. Due are to the contributing long-term meltingto this [6 of]. the glacier, the compressed rock is gradually risingThe on themapping outskirts of Greenland of Greenland. began According as early as to the scientific 15th century studies,, when the Greenland parts of the coast coast riseswere by mapped. 2.5 cm per However, year [3– 5systematic]. However, mapping it is also became scientifically possible confirmed only in that the some 20th partscentury of , theusing glacier aerial are photogrammetry even increasing. and This later information satellites. indicates A significant that the achievement condition ofwas the the Green- Dan- landish geodetic cap must expedition continue to from be carefully 1931 to studied.1934. Many Today, aerial scientific photographs satellites were in particular taken using are contributingthree Heinkel to seaplanes. this [6]. These photographs are the perfect source for monitoring the con- ditionThe of mapping Greenlandic of Greenland glaciers (Figure began as2). early as the 15th century, when parts of the coast were mapped.During the However, seventh systematicThule Expedition mapping, a systematic became possible survey only of the in southeast the 20th century, coast of usingGreenland aerial photogrammetrywas carried out from and 1932 later satellites.to 1933. Nowadays, A significant at the achievement Natural History was the Museum Danish geodeticof Denmark, expedition there is from a long 1931-term to 1934. project Many called aerial “AirBase photographs”. It is werea database taken usingwhich threehas a Heinkelquarter seaplanes.of a million These aerial photographs photos recorded are the by perfect Danish source survey for agencies monitoring in the the period condition from of1930 Greenlandic to the 1980s. glaciers This (Figureis a unique2). source of information about glaciers [7,8].

FigureFigure 2. 2.The The Knud Knud Rasmussen Rasmussen Glacier Glacier in in a photoa photo from from the the Danish Danish Greenland Greenland expedition expedition (the (the seventh sev- Thuleenth Thule Expedition Expedition 1932–1933); 1932–1933); (photo: (photo: the Arctic the Arctic Institute, Institute, https://arktiskinstitut.dk https://arktiskinstitut.dk).).

2. ProjectDuring Aims the seventh Thule Expedition, a systematic survey of the southeast coast of GreenlandBased was on the carried scope out of fromthis Special 1932 to Issue 1933. call Nowadays, “Analyses at thein Geoma Naturaltics: History Processing Museum Spa- oftial Denmark, Data on History there is aand long-term Today”,project we focused called on “AirBase”. the processing It is a of database spatial and which historical has a quarterdata. Two of a goals million have aerial been photos defined recorded as follows: by Danish survey agencies in the period from 1930 to the 1980s. This is a unique source of information about glaciers [7,8]. 2.1. Measuring the Glacier Flow Speed 2. Project Aims In this project, the primary aim was to calculate the speed of movement and its dis- tributionBased of on the the glacier scope ofbased this Specialon two Issuedrone call over “Analysesflights performed in Geomatics: at different Processing time Spatials. This Data on History and Today”, we focused on the processing of spatial and historical data. method compares two points clouds derived from photo sets to get these parameters. Two goals have been defined as follows:

2.1. Measuring the Glacier Flow Speed In this project, the primary aim was to calculate the speed of movement and its distribution of the glacier based on two drone overflights performed at different times. This method compares two points clouds derived from photo sets to get these parameters. Appl. Sci. 2021, 11, x FOR PEER REVIEW 3 of 19

Appl. Sci. 2021, 11, 754 3 of 19 2.2. Detection of Long-Term Changes in the Extent of the Glacier Another goal was to detect changes in the extent of the glacier based on the analysis of2.2. historical Detection image of Long-Term data. The Changes on-site in measurement the Extent of the in Glacier 2019 was compared with historical imageAnother data. There goal are was freely to detect available changes data in from the extentsatellites of theon the glacier web, based accessed on the via analysis Google Earth,of historical new free image satellite data. data The on-sitefrom the measurement Copernicus insystem 2019, was and compared we also managed with historical to get historicalimage data. photographic There are freely aerial available data from data the from Danish satellites archive. on theThe web, data accessedcan be processed via Google in geographiEarth, newc freeinformation satellite systemdata from (GIS the), and Copernicus the glacier system, extent and changes we also over managed time can to get be modelledhistorical. photographic aerial data from the Danish archive. The data can be processed in geographic information system (GIS), and the glacier extent changes over time can be 3.modelled. Study Area Based on the collaboration between the Beuth Hochschule für Technik Berlin, TU 3. Study Area Cottbus, and the Czech Technical University in Prague, Faculty of Civil Engineering, and the successfullyBased on the first collaboration joint expedition between in 2015, the a Beuthsecond Hochschule expedition fürwas Technik carried out Berlin, in 2019 TU [9]Cottbus,. This andexpedition the Czech with Technical the research University ship Dagmar in Prague, Aaen Faculty led by of CivilMr. Arved Engineering, Fuchs was and plannedthe successfully along the first east joint coast expedition of Greenland in 2015,. Our a second part in expedition the expedition was carried was photogramme- out in 2019 [9]. tryThis. expedition with the research ship Dagmar Aaen led by Mr. Arved Fuchs was planned alongOur the goals east coast of the of photogrammetrical Greenland. Our part works in the during expedition the wasGreenland photogrammetry. expedition 2019 was toOur create goals a of detailed the photogrammetrical map of the abandoned works during U.S. military the Greenland air base expedition Bluei East 2019 II wasand researchto create athe detailed movement map of the abandonedKnud Rasmussen U.S. military Glaciers air using base Blueithe drone. East II Both and researchsites are locatedthe movement within a of few the days Knud of Rasmussen sailing from Glaciers Kulusuk/ usingTasiilaq the drone.(Figure Both 3). This sites research are located is a within a few days of sailing from Kulusuk/Tasiilaq (Figure3). This research is a follow- follow-up to a number of projects focused on documenting glaciers using drones and de- up to a number of projects focused on documenting glaciers using drones and detecting tecting changes using point clouds [10–12]. As a study area analyzed in this article, the changes using point clouds [10–12]. As a study area analyzed in this article, the famous famous Knud Rasmussen Glacier was selected. It is a well-known formation and can be Knud Rasmussen Glacier was selected. It is a well-known formation and can be reached reached by sea (Figures 3 and 4). The choice of this glacier was given mainly by the pos- by sea (Figures3 and4). The choice of this glacier was given mainly by the possibility to sibility to obtain archival aerial data photographed during the Danish Greenland expedi- obtain archival aerial data photographed during the Danish Greenland expedition in 1930s. tion in 1930s.

(a) (b)

Figure 3. 3.(a ) Map(a) of Map Greenland of Greenland (http://www.getamap.net/maps/greenland_%5B_denmark_%5 (http://www.getamap.net/maps/greenland_%5B_den- mark_%5DD/ostgronland/_knud_rasmussen_glacier//ostgronland/_knud_rasmussen_glacier/) and ()b ) theand Knud (b) Rasmussen the Knud Glacier Rasmussen (66.0817028N, Glacier (66.0817028N, 36.3402919W), (http://www.mapy.cz). 36.3402919W), (http://www.mapy.cz). Appl.Appl. Sci.Sci. 20212021,, 1111,, xx FORFOR PEERPEER REVIEWREVIEW 44 ofof 1919

Appl. Sci. 2021, 11, 754 4 of 19 Appl. Sci. 2021, 11, x FOR PEER REVIEW 4 of 19

Figure 4. Knud Rasmussen Glacier; it appears to be a small glacier, but its width is over 2 km. Figure 4.4. Knud Knud Rasmussen Rasmussen Glacier; Glacier; it appears it appears to be toa smallbe a smallglacier, glacier, but its widthbut its is width over 2 iskm. over 2 km. Figure 4. Knud Rasmussen Glacier; it appears to be a small glacier, but its width is over 2 km. 4. Used Instruments 4.4. UsedUsed InstrumentsInstruments RemotelyRRemotelyemotely piloted pilotedpilotedpiloted aircraft aircraft aircraft aircraft system system system system (RPAS (RPAS)((RPASRPAS) is a)) is lessis is a a a lessused less less used usedbut used correct but but but correct correct correct acronym acronym acronym acronym for a for for for a a a drone [13]. The eBee drone, senseFly, from Switzerland, was used in this case for different dronedronedrone [ [113[133].]].. The TheThe eBeeeBeeeBee drone,drone,drone, senseFly, senseFly,senseFly, from fromfrom Switzerland, Switzerland,Switzerland,was waswasused usedused in inin this thisthis case casecase for forfor different differentdifferent reasons. (1) It is light, made from Styrofoam, easy to use, takes off by hand, and lands reasons.reasons.reasons. (1)((1)1) ItIt isisis light,light,light, mademademade fromfromfrom Styrofoam,SStyrofoam,tyrofoam, easyeasyeasy tototo use,use,use, takestakestakes offoffoff by byby hand, hand,hand, and andand lands landslands witwithouthout any anyany special specialspecial landing landinglanding pad, pad,pad, (2) (it(2)2) is it iteasy isis easyeasy to transport toto transporttransport (in a padded (in(in aa paddedpadded backpack), backpack),backpack), can be cancan bebe withoutdivided anyinto specialseveral parts landing for transport, pad, (2) it is is electrically easy to transport powered, (in and a uses padded small backpack), batteries, can be divideddivided intointo severalseveral partsparts forfor transport,transport, isis electricallyelectrically powered,powered, andand usesuses smallsmall batteries,batteries, divided(3) it has into a relatively several long parts time for of transport, flight up to is electrically40 min and altitude powered, acce andss of uses hundreds small of batteries, (3) it has a relatively long time of flight up to 40 min and altitude access of hundreds of (3)(m3), it becauseit has has a arelatively it relatively is a winged long long drone, time time and of of flight(4) flight it flies up up towith to 40 full40 min minautonomy and and altitude altitude based access on acce GNSS ofsshundreds (Globalof hundreds of m, of becausemmNavigation,, becausebecause it is Satellite itit a isis winged aa wingedwinged System) drone, drone,drone, and and IMU andand (4) (Inertial ((4) it4) fliesitit fliesflies Measurement with withwith full fullfull autonomy autonomyautonomy Unit) devices based basedbased and on ononthe GNSS GNSS GNSSflight (Global (Global(Global NavigationNavigationNavigationplan, and has Satellite SatelliteSatellite changeable System) System)System) cameras and andand (Figure IMU IMUIMUs (Inertial (Inertial (Inertial5 and 6). Measurement MeasurementHowever,Measurement the drone Unit) Unit)Unit) weighs devices devicesdevices less and andand than the thethe flight flightflight plan,plan,plan,one kilogram and andand has hashas and changeable changeablechangeable the flight cameras camerascamerascan be affected (Figures ((FigureFigure bys5s 5gusts5and andand6 ). of 6).6). However,wind, However,However, and thefurthermore, tthehe drone dronedrone weighs weighsweighscameras less less less than thanthan oneoneoneused kilogram kilogramkilogram in this type andandand is low thethethe-cost flight flightflight only. can cancan Unfortunately, be bebe affected affectedaffected by bytheby gusts guststypegusts used of ofof wind, wind, wind,was not and andand a GNSS furthermore, furthermore,furthermore, RTK (Real cameras camerascameras usedusedusedTime ininKinematic) this this typetype type system, isis is lowlow low-cost- -whichcostcost only.only. allows only. Unfortunately,Unfortunately, Unfortunately,precise georeferencing thethe the typetype type of usedused outputs used wwasas without was notnot a nota GNSSGNSS geodet- a GNSS RTKRTK (RealRTK(Real (RealTimeTimeically TimeKinematic)measuredKinematic) Kinematic) ground system,system, control system, whichwhich points which allowsallows (GCP). allows precisprecis ee precise georeferencinggeoreferencing georeferencing ofof outputsoutputs of outputs withoutwithout without geodet-geodet- geodeticallyicallyically measuredmeasured measured groundground ground controlcontrol control pointspointspoints (GCP).(GCP). (GCP).

Figure 5. The Remotely piloted aircraft system (RPAS) eBee.

FigureFigureFigure 5. 5.5. The TheThe Remotely RRemotelyemotely piloted pilotedpiloted aircraft aircraftaircraft system systemsystem (RPAS) ((RPASRPAS))eBee. eBeeeBee..

(a) (b) Figure 6. (a) Spectral response of the Canon PowerShot ELPH 110 HS (NIR) and (b) the changeable cameras for the eBee drone: near-infrared (NIR), red-edge (RE), and red-green-blue (RGB). The cam- era has a 1/1.7″ back-illuminated CMOS (Complementary metal–oxide–semiconductor) sensor, im- age size 12.1 MPix,(( acameraa)) weight 153g excl. battery. ((bb)) Figure 6. (a) Spectral response of the Canon PowerShot ELPH 110 HS (NIR) and (b) the changeable FigureFigure 6. 6.( a(a)) Spectral Spectral response response of of the the Canon Canon PowerShot PowerShot ELPH ELPH 110 110 HS HS (NIR) (NIR) and and ( b(b)) the the changeable changeable cameras for the eBee drone: near-infrared (NIR), red-edge (RE), and red-green-blue (RGB). The cam- camerascameras forfor the eBee eBee drone drone:: near near-infrared-infrared (NIR (NIR),), red red-edge-edge (RE (RE),), and and red red-green-blue-green-blue (RGB). (RGB). The Thecam- eraera hashas aa 1/1.71/1.7″″ backback--illuminatedilluminated CMOSCMOS ((ComplementaryComplementary metalmetal––oxideoxide––semiconductor)semiconductor) sensor,sensor, im-im- camera has a 1/1.7” back-illuminated CMOS (Complementary metal–oxide–semiconductor) sensor, ageage sizesize 12.112.1 MMPPix,ix, cameracamera weightweight 153g153g excl.excl. batterbatteryy.. image size 12.1 MPix, camera weight 153g excl. battery. Appl. Sci. 2021, 11, x FOR PEER REVIEW 5 of 19

5. Data Capturing Image data acquisition is described here. Data from RPAS were obtained directly at the research site, archival satellite data were obtained from free sources on the web and we also got unique historical aerial photos thanks to the kind approach of the Danish side.

5.1. RPAS Data Capturing Before the flights, flight planning is necessary. The eMotion software is used with the eBee system running on a notebook. This software uses a common GNSS device for navi- gation. During the flight planning in this software, satellite images from ordinary free web sources (Google, Microsoft, Nokia, and others) are used as a map background. Unfortu- nately, in Greenland, actual satellite data from these sources was unavailable, and it was not possible to precisely define the exact parameters and destination of the drone flight. This proved to be a problem with flight planning [10,13,14]. The front of the glacier was hundreds of m away from the latest satellite images, so the flight plan had to be estimated on-site (Figure 7). One single day was set aside for the measurement, and there was not more time. Two Appl. Sci. 2021, 11, 754 changeable cameras were at the drone’s disposal during the expedition. Unfortunately,5 of 19 due to a malfunction in the RGB camera, only the NIR (near-infrared) camera was used. As it turned out, the NIR images were more suitable for capturing ice than an RGB camera, because the glacier ice is off-white to light blue, which is not perfect for the image corre- 5.lation Data used Capturing for image processing. Using the drone eBee, two overflights of the same area wereImage performed data acquisition on 16 July is 2019 described with a here.time Dataspan fromof about RPAS four were hours. obtained During directly the first at theflight, research a total site, of archival162 NIR satelliteimages datawere were taken obtained (Figure from8). During free sources the second on the flight, web and170 NIR we alsoimages got uniquewere recorded. historical During aerial photosthe first thanks flight,to the the signal kind was approach lost due of theto a Danish very high side. rock cliff near the last flight line; the drone emergency system interrupted the flight and re- 5.1. RPAS Data Capturing turned it to the take-off point. For this reason, the number of images is not the same for bothBefore flights; the however, flights, the flight loss planning of eight images is necessary. did not The affect eMotion the project. software The islast used flight with line thewas eBee not systemfully completed; running on there a notebook. was no time This to software repeat the uses flight a common and therefore GNSS devicethe second for navigation.flight was slight Duringly modified. the flight planning in this software, satellite images from ordinary free webThe sources flight time (Google, took approximately Microsoft, Nokia, 35 min and at others) an altitude are used of 180 as m. a mapDue background.to the smaller Unfortunately,capacity of the in batteries Greenland, in actualthe frost, satellite we did data not from choose these a sources lower flight was unavailable, or a longer andflight. it wasAnother not possible problem to preciselywas the high define rocks the on exact both parameters sides of the and glacier. destination of the drone flight. This provedThe aim to was be a to problem record withboth flightshores planning of the moraine [10,13,14 (see]. The Figure front 7) of, which the glacier could was not hundredschange in of a mfew away hours, from unlike the latest the movement satellite images, of the glacier, so the flight which plan was had supposed to be estimated to be the on-site (Figure7). largest somewhere in the middle.

FigureFigure 7. 7.Flight Flight no.1,no.1, senseFlysenseFly eMotion eMotion 2 2 software software;; the the map map background background from from Google Google Earth Earth was was out of date (2016). The flight plan was prepared at the university before expedition, but the actual posi- out of date (2016). The flight plan was prepared at the university before expedition, but the actual tion of the glacier front was elsewhere. There was no internet on site, the flight plan had to be esti- position of the glacier front was elsewhere. There was no internet on site, the flight plan had to be mated and changed at the measurement site. estimated and changed at the measurement site.

One single day was set aside for the measurement, and there was not more time. Two changeable cameras were at the drone’s disposal during the expedition. Unfortunately, due to a malfunction in the RGB camera, only the NIR (near-infrared) camera was used. As it turned out, the NIR images were more suitable for capturing ice than an RGB camera, because the glacier ice is off-white to light blue, which is not perfect for the image correlation used for image processing. Using the drone eBee, two overflights of the same area were performed on 16 July 2019 with a time span of about four hours. During the first flight, a total of 162 NIR images were taken (Figure8). During the second flight, 170 NIR images were recorded. During the first flight, the signal was lost due to a very high rock cliff near the last flight line; the drone emergency system interrupted the flight and returned it to the take-off point. For this reason, the number of images is not the same for both flights; however, the loss of eight images did not affect the project. The last flight line was not fully completed; there was no time to repeat the flight and therefore the second flight was slightly modified. Appl. Sci. 2021, 11, 754 6 of 19 Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 19

FigureFigure 8. 8.A A typical typical photo photo using using NIR NIR camera camera Canon Canon PowerShot PowerShot ELPH ELPH 110 110 HS HS (NIR). (NIR).

5.2. TheHistorical flight Image timetook Dataapproximately 35 min at an altitude of 180 m. Due to the smaller capacityIn this of the part batteries of the project in the, the frost, aim we was did to not show choose the possibility a lower flight of using or a different longer flight. types Anotherof image problem data for was long the-term high monitoring rocks on both of the sides state of of the the glacier. glacier and especially the new processingThe aim of was historical to record aerial both photographs shores of theinto moraine an orthophoto, (see Figure which7), is which compatible could notwith changemodern in data. a few hours, unlike the movement of the glacier, which was supposed to be the largestThere somewhere is aerial in and the satellite middle. data available. Regular satellite imagery has been for ci- vilian use since the 1980s. Historically, this is mainly data from Landsat satellites. Newer 5.2. Historical Image Data satellite data can be found from the Copernicus system. Here we were interested in freely availableIn this data part only. of the These project, are theavailable aim was on toGoogle show Earth the possibility or on similar of using map different servers, typeswhere ofthey image are dataalready for processed long-term into monitoring a map form of the and state are ofof thelower glacier quality, and which especially doesthe notnew mat- processingter in this case of historical, because aerial only the photographs changing shape into an of orthophoto, the glacier front which was is compatibleanalyzed. Profes- with modernsional original data. satellite data acquired from public funds are today on data archive acces- sibleThere via access is aerial hub and like satellite http://scihub.copernicus.eu data available. Regular or satellitehttps://earthexplorer.usgs.gov/. imagery has been for civilian use since the 1980s. Historically, this is mainly data from Landsat satellites. Newer Aerial image data are often used for business, mainly for mapping and typically are not satellite data can be found from the Copernicus system. Here we were interested in freely free available on web. Really old historical aerial data can be found in museum data ar- available data only. These are available on Google Earth or on similar map servers, where chives. The Knud Rasmussen Glacier area was photographed during the Danish Green- they are already processed into a map form and are of lower quality, which does not land expedition by the seventh Thule Expedition from 1932 to 1933. It was thus possible matter in this case, because only the changing shape of the glacier front was analyzed. to study the condition of the glacier now and compare it with its condition almost ninety Professional original satellite data acquired from public funds are today on data archive years ago. Thank to Anders Anker Bjørk from the Department of Geoscience and Natural accessible via access hub like http://scihub.copernicus.eu or https://earthexplorer.usgs. Resource Management, University of Copenhagen, we got a set of historical aerial images gov/. Aerial image data are often used for business, mainly for mapping and typically from the Danish Greenland expedition in the 1930s (Figure 9). are not free available on web. Really old historical aerial data can be found in museum data archives. The Knud Rasmussen Glacier area was photographed during the Danish Greenland expedition by the seventh Thule Expedition from 1932 to 1933. It was thus possible to study the condition of the glacier now and compare it with its condition almost ninety years ago. Thank to Anders Anker Bjørk from the Department of Geoscience and Natural Resource Management, University of Copenhagen, we got a set of historical aerial images from the Danish Greenland expedition in the 1930s (Figure9). Appl. Sci. 2021, 11, 754 7 of 19 Appl. Sci. 2021, 11, x FOR PEER REVIEW 7 of 19

FigureFigure 9.9. The Heinkel Heinkel seaplane seaplane with with open open cockpit cockpit for for three three persons persons (pilot, (pilot, radio radio-operator,-operator, and and photographer in the back (photo: the Arctic Institute, https://arktiskinstitut.dk). The taking of the photographer in the back (photo: the Arctic Institute, https://arktiskinstitut.dk). The taking of photos was carried out from a height of up to 4500 m at a temperature of −40 degrees Celsius fre- the photos was carried out from a height of up to 4500 m at a temperature of −40 degrees Celsius quently. Nowadays, these conditions are hardly imaginable. frequently. Nowadays, these conditions are hardly imaginable.

6.6. DataData ProcessingProcessing AsAs was was written, written, in in this this project project two two types types of of data data were were used. used. First, First, aerial aerial images images taken taken fromfrom a a low low altitude altitude by by the the RPASRPAS eBee eBee on-site, on-site, and and second, second, satellitesatellite andand aerialaerial archivearchive data.data.

6.1.6.1.RPAS RPAS Data Data Processing Processing TheThe modernmodern digitaldigital aerialaerial photogrammetry photogrammetry servesserves thethe classic classic outputs outputs in in the the form form of of orthophotoorthophoto andand digitaldigital surfacesurface modelmodel(DSM) (DSM)based basedon on a a point point cloud cloud that that was was calculated calculated usingusing image image correlationcorrelation technology,technology, called called image-based image-based modelling modelling and and rendering rendering (IBMR) (IBMR) oror structurestructure fromfrom motion motion(SfM). (SfM).This Thistechnology technology usesuses aa setsetof ofhighly highly overlappedoverlapped photos.photos. BasedBased onon computing computing of of a a sparse sparse point point cloud cloud from from photographs, photographs, the the internal internal and and external external parametersparameters ofof thethe chosen camera can can be be computed. computed. After After this, this, a dense a dense point point cloud cloud is cal- is calculated,culated, which which is issimilar similar to tothe the point point cloud cloud measured measured directly directly using using laser laser scanners. scanners.

6.1.1.6.1.1.Creating Creatingof of PointPoint Clouds Clouds TheThe datadata couldcould onlyonly bebe processedprocessed inin aa laboratory laboratory atat the the university university andand not not directly directly duringduring thethe expedition.expedition. TheThe calculationcalculation waswas performedperformed inin Pix4DPix4D andand later later in in Metashape Metashape software;software; both both software software work work with with parameters parameters from from GNSS GNSS and and IMU IMU devices devices on on board board the the eBee drone. Metashape software provides a more detailed point cloud, which is why these eBee drone. Metashape software provides a more detailed point cloud, which is why these outputs were used in the next processing. outputs were used in the next processing. A computing of the glacier shift measured from two points clouds is non-trivial, A computing of the glacier shift measured from two points clouds is non-trivial, be- because of the nonlinear movement of the glacier. The glacier moves slowly at the edges cause of the nonlinear movement of the glacier. The glacier moves slowly at the edges and and fastest in the middle. The difference or the discrepancies between two points clouds is fastest in the middle. The difference or the discrepancies between two points clouds is usually calculated in CloudCompare software. usually calculated in CloudCompare software. By the computing in Metashape software, 74,416 tie points were found in the first set By the computing in Metashape software, 74,416 tie points were found in the first set of images for a sparse point cloud. In the second set of images, the number of tie points of images for a sparse point cloud. In the second set of images, the number of tie points was 92,693. After the depth map was created, 221,341,562 points for a dense point cloud was 92,693. After the depth map was created, 221,341,562 points for a dense point cloud were created for the flight and 231,635,556 for the second flight. Both point clouds were exportedwere created without for the modifying flight and their 231,635,556 original coordinate for the second systems. flight. It Both means point that clouds the second were pointexported cloud without was not modifying transformed the intoir original the coordinate coordinate system systems. of the It firstmeans point that cloud. the second pointA cloud point was cloud, not orthophoto, transformed and into DSM the wascoordinate made from system each of photothe first flight point and cloud. exported to theA ArcGIS point cloud, software orthophoto, (Figures 10 and and DSM 11). was made from each photo flight and exported to the ArcGIS software (Figures 10 and 11). The absolute position error of point clouds and orthophotos is given by the onboard GNSS device, which was in this case a common GNSS working with code mode with a precision of 3–5 m. For this reason it was not possible to use originally geocoded data for Appl. Sci. 2021, 11, x FOR PEER REVIEW 8 of 19

computing of the glacier movement, which was expected in m only. However, the internal accuracy is much higher, and in this case reaches 20 cm according to the pixel size. The precision of points in the point cloud derived from the photographs depends on filtering Appl. Sci. 2021, 11, 754 8 of 19 procedures. Points computed from only two photos were deleted, such as points with errors more than three pixels on images.

FigureFigure 10.10. FlightFlight no.1no.1 andand flightflight no.2no.2 (orthophoto (orthophoto in in near near infrared infrared range). range) On. On the the second second orthophoto ortho- therephoto are there shadows, are shadows, which makewhich the make picture the picture more plastic. more plastic. Appl. Sci. 11 Appl.2021 Sci., 2021, 754, 11, x FOR PEER REVIEW 99 ofof 1919

FigureFigure 11. Flight 11. Flight no.1 n ando.1 and no.2—digital no.2—digital surface surface model model (DSM). (DSM). The absolute position error of point clouds and orthophotos is given by the onboard 6.1.2. Joining of Both Point Clouds GNSS device, which was in this case a common GNSS working with code mode with a precisionAccurately of 3–5 m. Forreferencing this reason both it point was notclouds possible from individual to use originally flights was geocoded not typical, data forbe- cause no terrestrial geodetic measurement was made. There were no precisely measured computing of the glacier movement, which was expected in m only. However, the internal accuracy is much higher, and in this case reaches 20 cm according to the pixel size. The precision of points in the point cloud derived from the photographs depends on filtering Appl. Sci. 2021, 11, 754 10 of 19

procedures. Points computed from only two photos were deleted, such as points with errors more than three pixels on images.

6.1.2. Joining of Both Point Clouds Accurately referencing both point clouds from individual flights was not typical, because no terrestrial geodetic measurement was made. There were no precisely measured control points (GCP); it would be impossible to place and measure them on the glacier. GNSS equipment on the drone has an accuracy of 3–5 m, which is insufficient for precise georeferencing of projection centers by taking of photos and for georeferencing of computed points in point cloud derived from these photos. The assumption of movement in a few hours was one to several m maximally. The RTK system was unavailable and it would not work here due to the short measurement time and conditions, and no reference station was at our disposal either. The precise joining of both point clouds from the two overflights to one system was made as follows: Appl. Sci. 2021, 11, x FOR PEER REVIEW 11 of 19 The first model was defined as a reference and the second was transformed into it. However, a problem occurred with the joining of the second point cloud with the control points detected at the edge of the model only. Finally,The softwarethis transformation performs the result joining was of done both with point the clouds RMS using= 0.67 them. M controlany parameters points. These affectedcontrol the points resulting were accuracy. found on The the main first model,ones were but the only pixel on edgessize of of 10 the cm model, and the in type a stone andmoraine. structure These of the found surface, points including were relativelythe possible stable actual and deformation were used asof fixedthe glacier points. over In the timemiddle (Equation of the (3 glacier,)). Better considerable results cannot movement be expected was under expected, the andcurrent it was conditions not possible and to chosendefine equipment. high-quality It should fixed controlbe noted points; that the due model to this, was the created calculation from using images the taken new bundlein a timeadjustment span of approximately process deformed 35 min the. Even second in this model short by time, correcting parts theof the elements glacier ofmoved external, whichorientation. affected the The result model. curled, then a deformation occurred, and it reached more than two m (Figure 12).

FigureFigure 12. Schematic 12. Schematic deformation deformation of both of both models models after after joining joining the point the point clouds clouds with with control control points, points, foundfound at the at theedge edge of the of themodel model in a instone a stone moraine moraine (it should (it should be stable) be stable).. By joining By joining both both point point clouds, clouds, the thesecond second was was deformed deformed in the in themiddle middle due due to the to therecalculation recalculation of the of thebundle bundle adjustment adjustment (e.g. (e.g.,, the the internalinternal and and external external parameters parameters can canbe changed). be changed). For Forthis thisreason, reason, we weused used both both point point clouds, clouds, as as theythey were were processed processed separately separately,, and and the thejoining joining was was made made using using similarity similarity transformation transformation only only in in external software. external software.

It was necessary to use a different1.001 procedure.0 −0.001 Models−85.045 (point clouds) cannot be joined only at fixed control points at the0 edges1. of001 the− model,0.001 because−90.371 then generally an unknown 푇푡푟푎푛푠 = ( ) deformation occurs in the middle0.001 of the0.001 model.1.001 Both point−0.034 clouds were calculated(3) sepa- rately, and the second point cloud0 was joined0 only0 by a similarity1 transformation (shifts, three rotations, scaling) using six푅푀푆 control= 0.667566 points[ directly푚] in CloudCompare software and university software Tran independently together [15]. It finally gave usable results with an Equation (3) shows the final RMS for both point clouds joining and the transfor- RMS (root mean square) of 0.67 m, 0.69 m, respectively. A total of eight potential control mation matrix (CloudCompare software).

6.1.3. Analyzing the Glacier Flow Speed CloudCompare software calculates the distance between two similar point clouds in the direction of normals; in this case, the solution cannot be successful due to irregular shifts of the glacier. Here we do not have two almost identical point clouds, between which we measure the differences in the perpendicular direction to the surfaces, meaning in direction of nor- mals. Here is the original point cloud and a new point cloud, which was created by irreg- ularly differentially moving the first point cloud. Measuring the movement of the flow and its direction is thus much more complicated. It is necessary to define a point in the original point cloud and, according to its unique surroundings, try to find this point in the second shifted point cloud. If we succeed, we have the direction and magnitude of the shift. These parameters will be different for each point. The M3C2 (Model-to-Model Cloud Comparison) distance in a special module in CloudCompare software solves this problem. The next part was the analysis of the joined models and the search for real shifts in time as described above. A comparison of changes was performed using the Model-to- Appl. Sci. 2021, 11, 754 11 of 19

points were found, and only six were used, which was an error less than twice of mean error (RMS). No exact control points measured with a precise GNSS device were at our disposal. A similarity transformation in 3D was used (Equation (1)):       X X0 x  Y  =  Y0  + m·R· y  (1) Z Z0 z

where (X, Y, Z) are the final coordinates in the system of first model, (X0, Y0, Z0) are the shifts, m is the scale and (x, y, z) are the coordinates in the second model. The spatial seven-element transformation—the similarity transformation—is given by seven unknowns, three translations, three rotations, and one scaling. It can be possible to use the affine spatial transformation, which similar to the seven-element transformation, but it has nine unknowns, and there are three scale values—a different scale for each axis. To calculate the transformation key, it is necessary to calculate approximate values and proceed with these by iteration. Both types of transformation were calculated, but the difference between them was minimal in our case, only in units of centimeters. However, in the case of incorrectly determined identical points, a problem with the convergence of the calculation occurs. Both types of transformation are programmed in Delphi, originally at the Faculty of Civil Engineering, such as other university software for geodetical using [15]. m is a scalar (similarity transformation) or M is a matrix (affine transformation, Equation (2):   mx 0 0 M =  0 my 0  (2) 0 0 mz Finally, this transformation result was done with the RMS = 0.67 m. Many parameters affected the resulting accuracy. The main ones were the pixel size of 10 cm and the type and structure of the surface, including the possible actual deformation of the glacier over time (Equation (3)). Better results cannot be expected under the current conditions and chosen equipment. It should be noted that the model was created from images taken in a time span of approximately 35 min. Even in this short time, parts of the glacier moved, which affected the result.

 1.001 0 −0.001 −85.045   0 1.001 −0.001 −90.371  Ttrans =  RMS = 0.667566[m] (3)  0.001 0.001 1.001 −0.034  0 0 0 1

Equation (3) shows the final RMS for both point clouds joining and the transformation matrix (CloudCompare software).

6.1.3. Analyzing the Glacier Flow Speed CloudCompare software calculates the distance between two similar point clouds in the direction of normals; in this case, the solution cannot be successful due to irregular shifts of the glacier. Here we do not have two almost identical point clouds, between which we measure the differences in the perpendicular direction to the surfaces, meaning in direction of normals. Here is the original point cloud and a new point cloud, which was created by irregularly differentially moving the first point cloud. Measuring the movement of the flow and its direction is thus much more complicated. It is necessary to define a point in the original point cloud and, according to its unique surroundings, try to find this point in the second shifted point cloud. If we succeed, we have the direction and magnitude of the shift. These parameters will be different for each point. The M3C2 (Model-to-Model Cloud Comparison) distance in a special module in CloudCompare software solves this problem. Appl. Sci. 2021, 11, 754 12 of 19 Appl.Appl. Sci. Sci. 2021 2021, ,11 11, ,x x FOR FOR PEER PEER REVIEW REVIEW 1212 of of 19 19

The next part was the analysis of the joined models and the search for real shifts in time ModelModelas described CloudCloud above.ComparisonComparison A comparison (M3C2)(M3C2) method ofmethod changes [[116]6] was.. ThisThis performed isis aa methodmethod using ofof thecalculatingcalculating Model-to-Model thethe dis-dis- tancetanceCloud oror Comparison discrediscrepanciespancies (M3C2) betweenbetween method twotwo cloudsclouds [16]. This generally,generally, is a method andand notnot of inin calculating thethe classicalclassical the directiondirection distance ofof or normalsnormalsdiscrepancies.. TheThe methodmethod between calculatescalculates two clouds distancesdistances generally, onlyonly and forfor not soso-- incalledcalled the classical corecore pointspoints direction,, whichwhich of isis normals. thethe se-se- lectionlectionThe method ofof aa smallersmaller calculates samplesample distances ofof pointspoints only fromfrom for thethe so-called wholewhole reference corereference points, pointpoint which cloudcloud is.. the TheThe selection pointspoints areare of usuallyusuallya smaller selectedselected sample basedbased of points onon thethe from minimumminimum the whole distancedistance reference theythey point mustmust cloud. contain.contain. The InIn pointsthethe firstfirst are step,step, usually thethe normalnormalselected vectorsvectors based forfor on each theeach minimum pointpoint areare calculated. distancecalculated. they AllAll must pointspoints contain. locatedlocated In atat the thethe first maximummaximum step, the sphericalspherical normal distancedistancevectors for ((ss)) specifiedeachspecified point byby are thethe calculated. useruser fromfrom All thethe points corecore pointspoints located areare at usedused the maximum toto calculatecalculate spherical thethe normals.normals. distance IfIf thethe(s) cloudsspecifiedclouds containcontain by the normalnormal user from vectorsvectors the core fromfrom points postpost areprocessingprocessing used to, calculate, thesethese normalsnormals the normals. cancan bebe If used.used. theclouds containForFor each normaleach core core vectors point, point, froma a normal normal post vector vector processing, is is calculated, calculated, these normals which which is canis the the be average average used. of of the the normals normals ofof allall Forpointspoints each thatthat core lielie point, fromfrom athethe normal corecore point vectorpoint withwith is calculated, thethe maximummaximum which distancedistance is the averages.s. ThisThis of vectorvector the normals formsforms thetheof all axisaxis points ofof thethe that cylinder.cylinder. lie from In theIn thethe core nextnext point step,step, with aa cylinder thecylinder maximum withwith distances.aa useruser--defineddefined This vector heightheight forms ((hh)) andand the radiusradiusaxis of (( therr)) isis cylinder. interposedinterposed In the byby nextaa cloud.cloud. step, TheThe a cylinder corecore pointpoint with formsforms a user-defined thethe centcentee heightrr ofof thisthis (h ) cylinder.cylinder. and radius TheThe (r ) positionpositionis interposed of of a a point point by a is is cloud. calculated calculated The coreas as the the point average average forms of of theal alll points points center that that of this are are cylinder.in in this this cylinder. cylinder. The position A A more more of detaileddetaileda point isdescriptiondescription calculated ofof as thethe the methodmethod average cancan of beallbe foundpointsfound in thatin LLagueague arein etet this al.al. (2013)(2013) cylinder. [17[17–– A19]19] more,, ((FigureFigure detailedss 1133 andanddescription 1144)).. of the method can be found in Lague et al. (2013) [17–19], (Figures 13 and 14).

FigureFigureFigure 13.13. 13. HistogramHistogramHistogram ofof of thethe the ModelModel Model-to-Model--toto--ModelModel CloudCloud Cloud ComparisonComparison Comparison ((M3C2 (M3C2)M3C2)) distancedistance distance... NegativeNegative Negative valuesvalues values areareare errors errors errors from from from the the the computation computation computation occur occur occur on on on the the the borders borders borders only, only, only, caused caused caused by by by missing missing missing marginal marginal marginal data data data in in in both both both files.files.files. TheseThese These datadata data werewere were filteredfiltered filtered out.out. out.

FigureFigureFigure 14.14. AreasAreas of of significantsignificantsignificant change changechange on onon the thethe Knud KnudKnud Rasmussen RasmussenRasmussen Glacier GlacierGlacier in the inin timethethe timetime span spanspan of four ofof hours. fourfour hourshoursThe values:.. TheThe values:values: the flow thethe speed; flowflow thespeedspeed purple;; thethe colored ppurpleurple coloredpartscolored on parts theparts left onon are thethe caused leftleft areare by causedcaused missing byby data missingmissing from dataflightdata fromfromNo.1. flightflight NoNo..1.1. Appl.Appl. Sci. Sci. 20212021, 11, 11, x, 754FOR PEER REVIEW 1313 of of 19 19

TheThe results results of of the the analysis analysis show show that that the the glacier glacier moves moves fastest fastest in in the the middle middle with with the the speedspeed of of 2 2–3–3 m inin aa timetime span span of of four four hours. hours. The The histogram histogram (Figure (Figure 13) shows13) shows that thethat mean the meanspeed speed of the of glacier the glacier is one is meterone meter per fourper four hours. hours From. From the resultsthe results it can it can be concluded be concluded that thatwith with an erroran error of 0.67 of 0. m67 the m glacier the glacier Knud Knud Rasmussen Rasmussen moves moves at an averageat an average speed ofspeed around of aroundsix m per six day.m per The day. fastest The parts fastest can parts have can a speed have of a 12–15speed m of per 12– day,15 m while per theday, border while partsthe bordermove parts practically move practically by this method by this immeasurably method immeasurably in the short in the time short span time(Figure span 14 (Figure). This 14agrees). This with agrees other with studies other [studies10,20–22 [10,20]. –22].

6.2.6.2. Processing Processing of of Historical Historical Image Image Data Data 6.2.1.6.2.1. Historical Historical Aerial Aerial Images Images from from the the 1930s 1930s TheThe set set consists consists of of several several oblique oblique aerial aerial images images in in the the form form of of scanned scanned photocopies. photocopies. TheThe original original images images were were taken taken by by the the Fairchild Fairchild F F-8-8 photogrammetric photogrammetric camera, camera, which which was was releasedreleased in in 1930. 1930. Original Original images images from from the theFairchild Fairchild F-8 camera F-8 camera were were taken taken on 5″ on × 7 5”″ film.× 7” Thefilm. focal The length focal lengthwas 240 was mm 240 (12 mm″). We (12”). found We foundonly this only information this information about aboutthe photos the photos.. The fiducialThe fiducial marks marks were werevery difficult very difficult to find, to find,and the and frame the frame data datawere wereunusable. unusable. It was It un- was certainuncertain whether whether the thephotographs photographs were were complete complete in inthe the original original format. format. For Fortunately,tunately, the the centcenterer of of the the image image was was highlighted highlighted by by a apuncture puncture and and a a mark mark in in the the photographs. photographs. All All scannedscanned paper paper photocopies photocopies had had to to be betransformed transformed into intodetected detected centers, cent fitteders, and fitted cropped and croppedaccording according to poorly to visiblepoorly fiducialvisible fiducial marks to marks the same to the format. same Projectiveformat. Proj transformationective trans- formationwas used, was which used, got which the best got results the best (Figure results 15 ().Figure 15).

(a) (b)

FigureFigure 15. 15. (a()a T) Thehe Fairchild Fairchild F F-8-8 camera, camera, (b (b) )a a historical historical photo photo with with the the pilot pilot and and photographer photographer in in the thebook. book. The The photographer photographer holding holding a Fairchilda Fairchild F-8 F camera-8 camera [7]. [7].

OnlyOnly eight eight historical historical images images were were selected, selected, from from which which a project a project was created was created to pro- to cessprocess image image information information into a into 3D amodel. 3D model. It was It necessary was necessary to define to define at least at leastthe basic the basic pa- rametersparameters instead instead of the of theelements elements of ofinternal internal orientation. orientation. A Afocal focal length length of of 240 240 mm mm wa wass used,used, and and the the pixel pixel size size (14.2 (14.2 µµm)m) was was derived derived from from the the size size of of the the scanned scanned photographs photographs inin pixels pixels and and lines lines based based on on the the original original image image size size of of5″ 5”× 5×″. T5”.he Theadditional additional part partof the of filmthe was film used was for used frame for information frame information,, such as suchtime and as time photo and number photo. Metash number.ape Metashape software wassoftware used. Automatic was used. processing Automatic was processing not possible was due not to possible the low duequality tothe of scanned low quality paper of photocopiesscanned paper and photocopiesthe unknownand internal the unknownand external internal parameters and external of the chosen parameters photogram- of the metricalchosen photogrammetricalimages. It was necessary images. to It find was suitable necessary tie to points find suitable manually; tie points in the manually;end, 25 tie in pointsthe end, were 25 used tie points for a werecorrect used calculation. for a correct A sparse calculation. cloud was A sparse calculated, cloud and was the calculated, points wereand filtered the points using were a “gradual filtered usingselection a “gradual” procedure selection”, which procedure,filtered out whichpoints filteredwith error out ofpoints bigger with than error ten pixels. of bigger This than parameter ten pixels. was Thisset empiri parametercally so was that set approximately empirically so 30% that ofapproximately all points were 30% deleted of all from points the were processing. deleted from The thecamera processing. parameters The camerawere recalculated parameters untilwere stable recalculated results were until obtained. stable results Some were pictures obtained. did not Some have pictures a regular did photogrammetric not have a regular overlapphotogrammetric; they were taken overlap; by hand they and were it takenwas alm byost hand always and a it pair was of almost similar always pictures. a pairFrom of thesimilar photo pictures.-flight sequence From the of photo-flight the Knud Rasmussen sequence of G thelacier, Knud all Rasmussenimages were Glacier, finally all success- images fullywere processed finally successfully. An illustrative processed. model of An the illustrative historical status model was of the calculated historical (Figure statuss 1 was6– 19calculated). (Figures 16–19). An interesting result is the fact that the face of the Knud Rasmussen Glacier changed little over time compared to the other, dramatically receding glaciers, such as Jakobshaven Glacier on the west coast [20,21], including right next to the mouth of the Karale Glacier (the glacier on the left part of the model). Appl. Sci. 2021,, 11,, 754x FOR PEER REVIEW 14 of 19 Appl. Sci. 2021, 11, x FOR PEER REVIEW 14 of 19

Appl. Sci. 2021, 11, x FOR PEER REVIEW 14 of 19

Figure 16. The created three three-dimensional-dimensional (3D)(3D) model of Karale (left) and Knud Rasmussen glacier mouths (right) with tie Figure 16. The created three-dimensional (3D) model of Karale (left) and Knud Rasmussen glacier mouths (right) with tie pointsFigure and 16. photo The created positions;positions three; 1.231.23-dimensional millionmillion faces.faces (3D) .model of Karale (left) and Knud Rasmussen glacier mouths (right) with tie pointspoints and and photo photo positions positions; 1.231.23 millionmillion faces faces. .

Figure 17. A detail of the 3D model—the Knud Rasmussen Glacier in the 1930s. Figure 17. A detail of the 3D model model—the—the Knud Rasmussen G Glacierlacier in the 1930s.1930s. Figure 17. A detail of the 3D model—the Knud Rasmussen Glacier in the 1930s.

Figure 18. An orthographic view of the 3D model shows glacier mouths; it can be joined with satel- lite and drone images. Unfortunately, the images were taken at a low angle oblique, which caused

a considerable amount of hidden parts due to the perspective (white areas). Figure 18. An orthographic view of the 3D model shows glacier mouths; it can be joined with satel- Figure 18. An orthographic view of the 3D model shows glacier mouths; it can be joined with satel- liteFigure and 18. droneAn orthographicimages. Unfortunately, view of the the 3D modelimages shows were taken glacier at mouths; a low angle it can oblique be joined, which with satellitecaused lite and drone images. Unfortunately, the images were taken at a low angle oblique, which caused aand considerabl drone images.e amount Unfortunately, of hidden parts the due images to the were perspective taken at a(white low angle areas) oblique,. which caused a a considerable amount of hidden parts due to the perspective (white areas). considerable amount of hidden parts due to the perspective (white areas). Appl. Sci. 2021, 11, 754 15 of 19 Appl. Sci. 2021, 11, x FOR PEER REVIEW 15 of 19

Appl. Sci. 2021, 11, x FOR PEER REVIEW 15 of 19

FigureFigure 19. 19.The The KnudKnud RasmussenRasmussen GlacierGlacier inin aa historicalhistorical photophoto (detail);(detail); unfortunately,unfortunately, from from this this photo, photo, itit is is not not possible possible to to get get 3D 3D information information (only (only the the glacier glacier delimitation). delimitation).

6.2.2.An Archive interesting Landsat result Satellite is the factData that the face of the Knud Rasmussen Glacier changed little overGoogle time Earth compared provides to the previews other, dramatically of Landsat satellite receding data glaciers, from such 1984 as to Jakobshaven 2016 (Figure FigureGlacier20) for 19. the on The theKarale Knud west Rasmussenand coast Knud [20 Glacier,Rasmussen21], including in a historical Glaciers. right photo nextThe (detail); data to the w mouthunfortunately,ere always of the taken from Karale on this the Glacier photo same, it(the dayis not glacier (directly possible on onto the getthe left 3D last part information day of of the the model). (only year, the 31 glacierDecember delimitation).), which is an advantage for compar- ison. In ArcGIS, all of the images were compared (Figure 20). It can be seen from the series 6.2.2.6.2.2.of images Archive Archive that Landsat Landsat major Satellitechanges Satellite Datato Data the front of the Knud Rasmussen Glacier did not occur untilGoogleGoogle after 20 Earth Earth11. The provides Karale previewspreviews Glacier began ofof LandsatLandsat to recede satellite satellite as early data data as from from 1990. 1984 1984 to to 2016 2016 (Figure (Figure 20) 2for0) for the the Karale Karale and and Knud Knud Rasmussen Rasmussen Glaciers. Glaciers. The The data data were were always always taken taken on theon samethe same day day(directly (directly on theon the last last day day of the of the year, year, 31 December),31 December which), which is an is advantagean advantage for for comparison. compar- ison.In ArcGIS, In ArcGIS all of, all the of imagesthe images were were compared compared (Figure (Figure 20). 2 It0). can It can be be seen seen from from the the series series of ofimages images that that major major changes changes to theto the front front of theof the Knud Knud Rasmussen Rasmussen Glacier Glacier did notdid occurnot occur until untilafter after 2011. 20 The11. KaraleThe Karale Glacier Glacier began began to recede to recede as early as early as 1990. as 1990.

(a) (b) (c) Figure 20. Landsat data from Google Earth (a) from 1984, (b) from 2002, (c) from 2016.

6.2.3. Archive Sentinel Satellite Data

As a new type of free satellite data, the Copernicus system was used, especially the (a) (b) (c) Sentinel 2a satellite (Figure 21). Unfortunately, in the archive, only data from last two year FigureFigure 20. 20. LandsatLandsatare on- line.data data fromOlder from Google Google data mustEarth Earth be( (aa) ) ordered.from from 1984, 1984, Most ( (bb)) from from of the 20 2002, 02,satellite ( (cc)) from from scenes 2016 2016. .are cloudy and cannot be used for this work, the images taken during summertime can be used without pro- 6.2.3.cessing Archive problems, Sentinel other Satellite can be Data affected by the shadows, which are in the Greenland due 6.2.3. Archive Sentinel Satellite Data to shortAs a daynew time type very of free long satellite (Figure data, 21b) .the We Copernicus processed s cenessystem from was s ummertimeused, especially and win-the SentineltertimeAs 2a abetween new satellite type 2019 (Figure of freeand 2 satellite20201). Unfortunately for data, prolonging the Copernicus, in theof our archive, study system only and was data research used, from especially olastf usability two year the of areSentineldifferent on-line. 2a image Older satellite data.data (Figure mustFor comparison be 21 ordered.). Unfortunately, withMost other of the in image satellite the archive, data, scenes only only are the datacloudy one from multispectral and last cannot two beyearchannel used are for on-line.(Red) this with work Older a ,pixel the data imagessize must of 10betaken m ordered. was during used. Most summertime These of the images satellite can were scenesbe used used arefor without cloudydelimitation pro- and cessingcannotof the problems,Knud be used Rasmussen for other this work,can Glacier be the affected front. images byAs taken theit is shadows, duringvisible on summertime which (Figure are 2 in1a,b can the), betheGreenland used glacier without lostdue a toprocessingfront short part day after problems,time summertime very long other (Figure can in 2020. be 2 affected1b) . We processed by the shadows, scenes from which summertime are in the Greenland and win- tertimedue to shortbetween day 2019 time and very 2020 long for (Figure prolonging 21b). We of processedour study scenes and research from summertime of usability and of differentwintertime image between data. 2019 For andcomparison 2020 for with prolonging other image of our studydata, only and researchthe one multispectral of usability of different image data. For comparison with other image data, only the one multispectral channel (Red) with a pixel size of 10 m was used. These images were used for delimitation of the Knud Rasmussen Glacier front. As it is visible on (Figure 21a,b), the glacier lost a front part after summertime in 2020. Appl. Sci. 2021, 11, 754 16 of 19

channel (Red) with a pixel size of 10 m was used. These images were used for delimitation Appl. Sci. 2021, 11, x FOR PEER REVIEWofthe Knud Rasmussen Glacier front. As it is visible on (Figure 21a,b), the glacier16 lost of 19 a front part after summertime in 2020.

(a)

(b)

FigureFigure 21. 21. SentinelSentinel 2a 2a data data;; ( (aa)) 31 31 July July 2020 2020,, ( (bb)) 9 9 October October 2020 2020,, lower lower part part of of the the glacier glacier is is affected affected byby hill hill shadows. shadows.

6.3.6.3. The The Knud Knud Rasmussen Rasmussen Glacier Glacier Time Time Changes AllAll collected collected image image data data w wereere processed processed in in the the ArcGIS ArcGIS software software to to the the final final map, map, which shows shows the the glacier glacier extent extent changes. changes. In In the the 1930s, 1930s, the the condition condition of the of the glacier glacier was was al- mostalmost the the same same as in as the in theearly early new new millennium millennium until until 2011. 2011. This Thisis quite is quite clearly clearly evident evident from thefrom georeferenced the georeferenced orthophoto orthophoto from the from 1930s the 1930scompared compared with the with Landsat the Landsat data from data from1984 to1984 201 to6. 2016.In 1932, In the 1932, north the– north–easterneastern part of partthe glacier of the front glacier was front located was at located a similar at aposition similar asposition it was asfrom it was 2011 from. The 2011. face Theof the face glacier of the from glacier 1932, from based 1932, on based the processing on the processing of the his- of toricalthe historical images, images, was behind was behind the position the position from from1984. 1984.However, However, while while we do we not do know not know ex- actlyexactly when when the the flight flight was was made, made, we we do do know know that that it it was was very very likely likely in in the the summer. summer. The question is is what the front of the glacier looked like in winter in the 1930s. Similarly, the datadata from from 2019 to 2020 are from a summer that was warm (Figures(Figures 22 andand 2323).). Landsat satellitesatellite dates dates are are always from winter; Sentinel 2a data are from summer and autumn. Appl.Appl. Sci. 2021,, 11,, 754x FOR PEERAppl. REVIEWSci. 2021, 11, x FOR PEER REVIEW 1717 of of 19 17 of 19

Figure 22. A final interpretationFigureFigure 22. of A 22.the final AKnud finalinterpretation Rasmussen interpretation of Glacier the of Knud the changes KnudRasmussen between Rasmussen Glacier 1932 Glacier andchanges 2020 changes between(2019 * between are 1932 data and 1932from 2020 and the (2019 2020 * are data from the eBee drone, some parts eBeewere(2019 drone, missing). * aresome data parts from were the eBeemissing). drone, some parts were missing).

Figure 23. The Knud RasmussenFigure 23. GlacierGlacier The Knud areaarea Rasmussen changeschanges between between Glacier 1932 1932 area and andchanges 2020; 2020; thebetween the basic basic state1932 state wasand was the2020; the basic state was yearthe year 1984 1984 (i.e., (i.e., the the state state “zero”).the “zero”). year 1984 Between Between (i.e., 1932the 1932 state and and “zero”). 1984, 1984, we weBetween do do not not have1932 have and data data 1984, about about we the the do glacier glaciernot have status. sta- data about the glacier sta- tus. tus. 7. Discussion and Conclusions 7. DiscussionThe first partand Conclusions of7. this Discussion research and looked Conclusions at the possibility of monitoring the flow rate and itsThe distribution first part of of this the researchThe face first of the looked part Knud of at this Rasmussenthe research possibility Glacierlooked of monitoring at in the eastern possibility the Greenland flow of ratemonitoring using and the flow rate and itsRPAS distribution based on pointof theits cloud face distribution of time the changes Knud of the Rasmussen detected face of fromthe Glacier Knud two overflights. inRasmussen eastern Greenland According Glacier in tousing eastern the Greenland using Appl. Sci. 2021, 11, 754 18 of 19

available results, the most significant changes occur in the middle part of the studied glacier. For computing the displacement of both point clouds derived from photo-sets in different time, the M3C2 distance from CloudCompare software was used. The displacement values histogram shows that the changes rarely exceed three m. A typical movement of the middle part of this glacier can be from 2 to 3 m in 4 h, but a mean speed based on the local histogram maximum is only 1 m in a time span of 4 h. There is a possible error of up to RMS 67 cm, given the inaccuracy of joining both point clouds. In conclusion, the use of a drone for monitoring the speed of the movement and its distribution of the glacier is possible even after a relatively short time span. The fastest running ones have a speed in the middle of up to tens of m per day, such as the Eqi Glacier or the Jakobshavn Glacier in west Greenland. In the case of the Knud Rasmussen Glacier, the flow rate in the middle nearby the glacier face is considerable and can reach a value of around 12–15 m per day at most; however, this is only an extrapolation, but it correlates with the measurement on other glaciers. Here, problems were found: only a short time span was used, only one measurement in the summertime was carried out, the parameters of the flight were not ideal due to the local conditions and lack of time. Any precisely measured control points were used. The problems with the precision and time of flight can be nowadays solved with a modern eBee RTK with a base station and with better quality batteries, and with more time for these experiments. The second goal of our research was aimed at an analysis of using and joining different image data for the monitoring of this glacier, mainly new processing of old historical aerial oblique images from the 1930s. From the historical images, we can deduce that the Knud Rasmussen Glacier receded inland mainly after 2011. This glacier retreat inland is not so progressive as that of other glaciers, which have receded km inland due to climate change. An analysis of historical aerial images, Landsat and Sentinel satellite images, and new drone orthophoto shows that the front of the Knud Rasmussen Glacier was in practically the same place until 2011. Only later did the glacier begin to recede. Unfortunately, data from 1933 to 1983 is unreachable or missing for us. We could not find out what the situation was in the fifty missing years. The new drone images, such as Sentinel data from 2019 to 2020, show the accelerated retreat of the glacier front in recent years. However, the newest (2019–2020) and the oldest (1932–1933) image data were obtained in the summer, satellite data from Landsat were always acquired on the last day of the year, and the cloud-free data from the Sentinel satellite were taken partially in summertime and partially in the autumn. Thus, a seasonal change is possible in this summer data, and the data have less significance with regard to long-term development. Even so, it is certain that the melting of the glacier is happening, and it is accelerating.

Author Contributions: Conceptualization, K.P.; methodology, K.P. and K.P.J.; technical advisory and supervision, K.P. and J.Š.; data processing, K.P., K.P.J., and J.Š.; investigation, K.P.; writing—original draft preparation, K.P.; writing—review and editing, K.P.J. and J.Š.; visualization, K.P., K.P.J., and J.Š.; project administration, K.P. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Czech Technical University in Prague Grant number SGS20/053/OHK1/1T/11 and the APC was funded by the Czech Technical University in Prague, FCE, department of Geomatics. Institutional Review Board Statement: No applicable. Informed Consent Statement: No applicable. Data Availability Statement: The used data were collected during the expedition and from open sources, as written (Landsat, museum data, Copernicus). Acknowledgments: We would like to thank Arved Fuchs and crew member Dagmar Aaen for an incredible experience and perfect service, to Robert Peroni (Tasiilaq, Greenland) for friendly welcome and valuable advice, and to Anders Anker Bjørk (University of Copenhagen) for the photographs provided and the immediate and friendly response. Appl. Sci. 2021, 11, 754 19 of 19

Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References 1. Greenland Rapidly Rising as Ice Melt Continues. 2020. Available online: http://www.miami.edu/index.php/news/releases/ greenland_rapidly_rising_as_ice_melt_continues/ (accessed on 5 October 2020). 2. Encyclopædia Britannica. 1999 Multimedia Edition. Available online: https://www.britannica.com/place/Greenland (accessed on 28 October 2020). 3. Stocker, T.F.; Qin, D.; Plattner, G.K.; Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.; Midgley, P.M.; et al. Climate Change 2013—The Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK, 2014. [CrossRef] 4. Slater, T.; Hogg, A.E.; Mottram, R. Ice-sheet losses track high-end sea-level rise projections. Nat. Clim. Chang. 2020, 10, 879–881. [CrossRef] 5. Mouginot, J.; Rignot, E.; Bjørk, A.A.; Van den Broeke, M.; Millan, R.; Morlighem, M.; Noël, B.; Scheuchl, B.; Wood, M. Forty-six years of Greenland Ice Sheet mass balance from 1972 to 2018. Proc. Natl. Acad. Sci. USA 2019, 116, 9239–9244. [CrossRef] 6. Sasgen, I.; Wouters, B.; Gardner, A.S.; King, M.D.; Tedesco, M.; Landerer, F.W.; Dahle, C.; Save, H.; Fettweis, X. Return to rapid ice loss in Greenland and record loss in 2019 detected by the GRACE-FO satellites. Commun. Earth Environ. 2020, 1, 1–8. [CrossRef] 7. Bjørk, A.A.; Kjær, K.H.; Larsen, N.K.; Kjeldsen, K.K.; Khan, S.A.; Funder, S.V.; Korsgaard, N.J. The Greenland Ice Sheet—80 Years of Climate Change Seen from the Air; Natural History Museum of Denmark, Faculty of Science, University of Copenhagen: Copenhagen, Denmark, 2014; 180p, ISBN 978-87-87519-46-5. 8. Bjørk, A.A.; Kjær, K.H.; Korsgaard, N.J.; Khan, S.A.; Kjeldsen, K.K.; Andresen, C.S.; Box, J.E.; Larsen, N.K.; Funder, S. An aerial view of 80 years of climate-related glacier fluctuations in southeast Greenland. Nat. Geosci. 2012, 5, 427–432. [CrossRef] 9. Näcke, L. Vergleich Kinematischer GNSS-Daten aus Ostgrönland (Comparison of Kinematic GNSS Data from East Greenland). Bachelor’s Thesis, Brandenburgische Technische Universität Cottbus-Senftenberg, Cottbus, Germany, 2020. 10. Pavelka, K.; Šedina, J.; Matoušková, E.; Hlaváˇcová, I.; Korth, W. Examples of different techniques for glaciers motion monitoring using InSAR and RPAS. Eur. J. Remote Sens. 2019, 52, 219–232. [CrossRef] 11. Bash, E.A.; Moorman, B.J.; Gunther, A. Detecting Short-Term Surface Melt on an Arctic Glacier Using UAV Surveys. Remote Sens. 2018, 10, 1547. [CrossRef] 12. Zwally, J.; Abdalati, W.; Herring, T.; Larson, K.; Saba, J.; Steffen, K. Surface Melt-Induced Acceleration of Greenland Ice-Sheet Flow. Science 2002, 297, 218–222. [CrossRef][PubMed] 13. Šedina, J.; Housarová, E.; Raeva, P. Using RPAS for the detection of archaeological objects using multispectral and thermal imaging. Eur. J. Remote Sens. 2019, 52 (Suppl. 1), 182–191. [CrossRef] 14. Korth, W.; Hoffman, U.; Hitziger, T.; Küchenmeister, T.; Pawelka, K. Klimabedingte Veränderungen des Inlandeises im südlichen Grönland. In 19. Internationale Geodätische Woche Obergurgl 2017; Herbert Wichmann Verlag: Berlin, Germany, 2017; ISBN 978-3-87907-624-6. 15. Štroner, M.; Pavelka, K. Linear spatial transformation. Geodetický a Kartografický Obzor 2001, 47, 233–235. 16. M3C2 (Plugin). CloudCompare. Available online: https://www.cloudcompare.org/doc/wiki/index.php?title=M3C2 (accessed on 10 October 2020). 17. Lague, D.; Brodu, N.; Leroux, J. Accurate 3D comparison of complex topography with terrestrial laser scanner: Application to the Rangitikei canyon (N-Z). ISPRS J. Photogramm. Remote Sens. 2013, 82, 10–26. [CrossRef] 18. Barnhart, T.B.; Crosby, B.T. Comparing Two Methods of Surface Change Detection on an Evolving Thermokarst Using High- Temporal-Frequency Terrestrial Laser Scanning, Selawik River, Alaska. Remote Sens. 2013, 5, 2813–2837. [CrossRef] 19. DiFrancesco, P.-M.; Bonneau, D.; Hutchinson, D.J. The Implications of M3C2 Projection Diameter on 3D Semi-Automated Rockfall Extraction from Sequential Terrestrial Laser Scanning Point Clouds. Remote Sens. 2020, 12, 1885. [CrossRef] 20. Gervaix, F. UAV Mapping of a Greenland Glacier. GIM J. 2018. Available online: https://www.gim-international.com/content/ article/uav-mapping-of-a-greenland-glacier (accessed on 28 October 2020). 21. Simonsen, S.B.; Sørensen, L.S. Implications of changing scattering properties on Greenland ice sheet volume change from Cryosat-2 altimetry. Remote Sens. Environ. 2017, 190, 207–216. [CrossRef] 22. Sandberg Sørensen, L.; Simonsen, S.B.; Forsberg, R.; Khvorostovsky, K.; Meister, R.; Engdahl, M.E. 25 years of elevation changes of the Greenland Ice Sheet from ERS, Envisat, and CryoSat-2 radar altimetry. Earth Planet. Sci. Lett. 2018, 495, 234–241. [CrossRef]