Spatiotemporal variations in the surface velocities of Antarctic Peninsula glaciers

36 Pages·2014·5.9 MB·English

TheCryosphereDiscuss.,8,5875–5910,2014 D
www.the-cryosphere-discuss.net/8/5875/2014/ isc TCD
u
doi:10.5194/tcd-8-5875-2014 s
s
©Author(s)2014.CCAttribution3.0License. io 8,5875–5910,2014
n
P
a
Thisdiscussionpaperis/hasbeenunderreviewforthejournalTheCryosphere(TC). p
e Spatiotemporal
PleaserefertothecorrespondingfinalpaperinTCifavailable. r
variations in the
|
surface velocities of
Spatiotemporal variations in the surface D
Antarctic Peninsula
is
c
u glaciers
velocities of Antarctic Peninsula glaciers s
s
io
n J.Chenetal.
P
1,2 1,2 1,2 a
J. Chen , C. Q. Ke , and Z. D. Shao p
e
r
1JiangsuProvincialKeyLaboratoryofGeographicInformationScienceandTechnology, TitlePage
|
NanjingUniversity,Nanjing,210023,China
Abstract Introduction
2KeyLaboratoryforSatelliteMappingTechnologyandApplicationsofStateAdministrationof D
is
Surveying,MappingandGeoinformationofChina,NanjingUniversity,Nanjing,210023,China c Conclusions References
u
s
s
Received:25October2014–Accepted:2November2014–Published:25November2014 io Tables Figures
n
Correspondenceto:C.Q.Ke([email protected]) P
a
p (cid:74) (cid:73)
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion. e
r
(cid:74) (cid:73)
|
D Back Close
is
c
u FullScreen/Esc
s
s
io
n
Printer-friendlyVersion
P
a
p
e InteractiveDiscussion
r
5875 |
Abstract D
is
c TCD
u
Velocity is an important parameter for the estimation of glacier mass balance, which s
s
directly signals the response of glaciers to climate change. Antarctic ice sheet move- io 8,5875–5910,2014
n
ment and the associated spatiotemporal velocity variations are of great significance to P
a
5 global sea level rise. In this study, we estimate Antarctic Peninsula glacier velocities pe Spatiotemporal
r
using the co-registration of optically sensed images and correlation (hereafter referred variations in the
to as COSI-Corr) based on moderate-resolution imaging spectroradiometer Level 1B | surface velocities of
data (hereafter referred to as MODIS L1B). The results show that the glaciers of Gra- D Antarctic Peninsula
ham Land and the Larsen Ice Shelf have substantially different velocity features. The iscu glaciers
s
10 Graham Land glaciers primarily flow from the peninsula ridge towards the Weddell sio
Sea and Bellingshausen Sea on the east and west sides, respectively. There are very n J.Chenetal.
largevelocityvariationsamongthedifferenticestreams,withaminimumof<20ma−1 Pa
and a maximum of 1500ma−1 (with an average of 100–150ma−1). Over the period pe
r
2000–2012, the glaciers of Graham Land accelerated in the south but slowed down in TitlePage
|
thenorth.Incontrast,theLarsenIceShelfflowsinarelativelyuniformdirection,mainly
15 Abstract Introduction
−1 D
towards the northeast into the Weddell Sea. Its average velocity is 750–800ma and is
the maximum is >1500ma−1. During the period 2000–2012, the Larsen Ice Shelf ex- cu Conclusions References
s
periencedsignificantacceleration.TheuseofCOSI-CorrbasedonMODISL1Bdatais s
io Tables Figures
suitable for glacier velocity monitoring on the Antarctic Peninsula over long time series n
P
and large spatial scales. This method is clearly advantageous for analysing macro- a
20 p (cid:74) (cid:73)
e
scale spatiotemporal variations in glacier movement. r
(cid:74) (cid:73)
|
1 Introduction D Back Close
is
c
u FullScreen/Esc
TheAntarcticicesheetiscloselyrelatedtosignificantproblems,suchasglobalclimate s
s
and eco-environment, as well as the future development of human society (Lucchitta ion
Printer-friendlyVersion
etal.,1993).Withtheintensificationofglobalwarming,glaciersareplayinganincreas- P
25 a
p
ingly evident role as an amplifier and indicator of global climate change (Shi and Liu, e InteractiveDiscussion
r
5876 |
2000).Glacialvelocityisanimportantfactorinfluencingiceflux(Rosenauetal.,2012) D
is
and an important parameter in the estimation of Antarctic ice sheet mass balance. c TCD
u
s
Additionally, glacial velocity directly signals the response of the Antarctic ice sheet to s
io 8,5875–5910,2014
global climate change. Thus, studies on the velocities of Antarctic glaciers have be- n
P
5 come important for developing global sea level rise models (Chen et al., 2007). a
p
Methods for monitoring Antarctic ice velocities are mainly classified into two cate- e Spatiotemporal
r
gories: in situ measurement and remote sensing monitoring. Stake measurement is variations in the
|
the first method of in situ measurement used for monitoring Antarctic ice velocities surface velocities of
D
(Hofmannetal.,1964;Dorreretal.,1969).Atpresent,globalpositioningsystem(here- Antarctic Peninsula
is
c
10 after referred to as GPS) technology has become the most important tool for field u glaciers
s
measurements of Antarctic glacier dynamics and ice surface terrain (Manson et al., s
io
2000). However, the harsh environment, high cost of field observation, and a relatively n J.Chenetal.
P
shortretestperiodhavesubstantiallylimitedinsitumeasurements(Urbini,2008;Sunil, a
p
e
2007). Remote sensing methods, including microwave and optical remote sensing, al- r
TitlePage
lowforrapiddevelopmentofAntarcticicevelocitymeasurementsusingsyntheticaper-
15 |
ture radar (hereafter referred to as SAR) data (Kimura et al., 2004; Dong et al., 2004; Abstract Introduction
D
Pattyn and Derauw, 2002). By means of interferometric SAR (hereafter referred to as is
c Conclusions References
InSAR) and feature tracking, researchers have obtained ice velocity measurements u
s
s
of the entire Antarctic ice sheet and generated an Antarctic ice sheet velocity profile io Tables Figures
n
(Rignotetal.,2011).SARdataareindependentofweatherconditions(cloudsandrain)
20 P
a
andhavehighaccuracy(Keetal.,2013).Nonetheless,thefrequentmeltingbehaviour p (cid:74) (cid:73)
e
of glaciers and snow on the Antarctic Peninsula is likely to cause a loss of coherence r
(cid:74) (cid:73)
in SAR images, accounting for the longer observation period of SAR feature tracking |
method than the InSAR method. Additionally, the impact of incidence angle in steep D Back Close
terrains can limit the visibility of some glaciers in SAR images, and a high-resolution is
25 c
u FullScreen/Esc
digital elevation model (hereafter referred to as DEM) is needed for terrain correction. s
s
Therefore, active microwave remote sensing is highly suitable for velocity monitoring io
n
Printer-friendlyVersion
over relatively short periods (Scheuchl et al., 2012). P
a
p
e InteractiveDiscussion
r
5877 |
Co-registration of optically sensed images and correlation (COSI-Corr) is a method- D
is
ology based on the calculation of grayscale pixel values using a simple algorithm. c TCD
u
s
COSI-Corr is advantageous because of its larger number of optical remote sensors, s
io 8,5875–5910,2014
longer time series, and rich data resources. It is therefore more suitable than SAR n
P
5 monitoring methods for estimating the speed of glacial movement and the analysis of a
p
relevantspatiotemporalvariationsoverlongtimeseries(Ferrignoetal.,1998;Liuetal., e Spatiotemporal
r
2012).BecausetheaccuracyofCOSI-Corrisprimarilydeterminedbythespatialreso- variations in the
|
lutionofapixel,Antarcticicesheetvelocitymonitoringbasedonopticalimagesmostly surface velocities of
D
adoptshigh-resolutionimages,suchasTM/ETM(Ferrignoetal.,1994),ASTER(Tiwari Antarctic Peninsula
is
10 et al., 2012), and SPOT (Ahn and Howat, 2010). Additionally, due to differences in cu glaciers
s
the attitudes of satellites, system errors occur during image splicing, and coordination s
io
registration errors between different images affect the accuracy of COSI-Corr. Thus, n J.Chenetal.
P
study on cross-correlation algorithms becomes the focus for improving the accuracy a
p
of COSI-Corr. For example, Heid and Kaab (2012) compared and evaluated six differ- er
TitlePage
ent matching algorithms: normalized cross-correlation operated in the spatial domain
15 |
(NCC), cross-correlation operated in the frequency domain using Fast Fourier Trans- Abstract Introduction
D
form (CCF), phase correlation operated in the frequency domain using Fast Fourier is
c Conclusions References
Transform(PC),cross-correlationoperatedinthefrequencydomainusingFastFourier u
s
s
Transformonorientationimages(CCF-O),phasecorrelationoperatedinthefrequency io Tables Figures
n
domainusingFastFourierTransformonorientationimages(PC-O)andthephasecor-
20 P
a
relationalgorithmusedinCOSI-Corr.Theuseofhigh-resolutionimagesformonitoring p (cid:74) (cid:73)
e
thevelocityofsingleicestreamsontheAntarcticPeninsulahasalsoresultedinimpor- r
(cid:74) (cid:73)
tantfindings.Forexample,thestudybyScambosetal.(2004)estimatedthevelocities |
of the Crane Glacier using TM images. As the collapse of the Antarctic ice shelf inten- D Back Close
sifies, great progress has been made in measuring ice shelf velocities (Yu et al., 2010; is
25 c
u FullScreen/Esc
Scheuchl et al., 2012). However, few studies have been reported for the Antarctic ice s
s
sheet with regard to ice velocities on larger spatial scales over longer time spans. io
n
Printer-friendlyVersion
AntarcticPeninsula glaciersarepolar continentalglaciers that havehigh surfaceve- P
a
locities (Khazendar et al., 2011). For these glaciers, MODIS data of 250m resolution pe InteractiveDiscussion
r
5878 |
meet the matching criteria of feature tracking. Additionally, medium-resolution images D
is
have the following advantages: (1) single images can cover a broader area, thereby c TCD
u
s
reducingerrorsincross-correlationscausedbydatasplicingandcoordinationregistra- s
io 8,5875–5910,2014
tion; and (2) the larger field of view allows for data coverage multiple times per day in n
P
5 thesameareaofthepolarregion,facilitatingtheselectionofhigh-qualitydatafromthe a
p
periodwithminimumcloud(Haugetal.,2010).Inthepresentstudy,weuseCOSI-Corr e Spatiotemporal
r
to estimate Antarctic Peninsula glacier surface velocities and analyse their spatiotem- variations in the
|
poral variations over the period 2000–2012 based on medium-resolution MODIS L1B surface velocities of
D
data. Antarctic Peninsula
is
c
u glaciers
s
s
10 2 Study area ion J.Chenetal.
P
a
The Antarctic Peninsula in West Antarctica (Fig. 1) is the largest peninsula of the p
e
◦ r
Antarcticcontinent,stretchingthefarthestnorthwardintothesea(63 S).Inthisregion,
TitlePage
the strata belong to a Cenozoic fold belt, and the bedrock is undulatory. The average |
Abstract Introduction
annualrainfallisupto500–600mm.Thepeninsula,alsoknownasMaritimeAntarctica, D
is
15 isthewarmestplacewiththemostsnowfallontheAntarcticcontinent.Overtheperiod cu Conclusions References
1952–2000, the Antarctic Peninsula experienced a temperature rise of 5.6◦C (Turner s
s
etal.,2005),andduring1955–1998,therewasaseasurfacetemperatureriseofmore ion Tables Figures
than 1◦C in the Pacific Ocean to the west of the peninsula (Meredith and King, 2005). P
a
p (cid:74) (cid:73)
Due to this temperature increase, the area lost from the Larsen Ice Shelf since 1986 e
r
2
20 hasexceeded8500km (PritchardandVaughan,2007).ThecollapseoftheLarsenIce (cid:74) (cid:73)
|
Shelf has caused substantial ice loss in upstream supplies of the Antarctic Peninsula,
e.g.upto6.8±0.3km3a−1 intheFlemingGlacier,whichacceleratedby50%relativeto D Back Close
is
the velocity in 1974 (Rignot et al., 2004). Pritchard and Vaughan (2007) monitored the cu FullScreen/Esc
s
velocities of more than 300 glaciers on the Antarctic Peninsula and found an average s
io
25 accelerationof12%between1992and2005,mainlyduetodownstreamglaciermass n Printer-friendlyVersion
P
loss.TheseauthorsfurtherestimatedthemassoutputoftheAntarcticPeninsulatobe a
p
60±46Gta−1 in 2005 (Pritchard and Vaughan, 2007). If the marine ice sheet in West e InteractiveDiscussion
r
5879 |
Antarctica completely disintegrates, sea level will rise 3.3m (Bamber et al., 2009). In D
is
short,AntarcticPeninsulaglaciersurfacevelocitieshavegreatimplicationsforestimat- c TCD
u
s
ing Antarctic ice sheet mass balance and the study of global climate change (Skvarca s
io 8,5875–5910,2014
et al., 1999; Osmanoglu et al., 2013). n
5 The Antarctic Peninsula has a number of affiliated islands that cover relatively small Pa
p
areas. In medium-resolution optical images, each of these affiliated islands occupies e Spatiotemporal
r
a very limited number of pixels, whereas the scope of the window set for cross- variations in the
|
correlation calculation is large relative to the island pixels. Thus, the affiliated islands surface velocities of
D
are neglected in this study to improve the accuracy of computation. Antarctic Peninsula
is
c
10 Duetofactorssuchastemperatureandoceancurrents,themeltlinesoficeshelves u glaciers
s
ontheAntarcticPeninsulavarywidelyduringdifferentyears(Fig.2).Ifsurfacefeatures s
io
of varying natures are used for feature tracking based on the digital number (hereafter n J.Chenetal.
P
referredtoasDN)valueofpixel,largeerrorswillariseinthecalculatedresults.Hence, a
p
we extract the minimum freeze/melt lines of sea ice during five different periods as the er
TitlePage
scope of cross-correlation calculation, mainly covering Graham Land and the Larsen
15 |
2
Ice Shelf (10.5km ). Abstract Introduction
D
is
c Conclusions References
u
s
3 Data s
io Tables Figures
n
TheuseofCOSI-Corrrequires:(1)atleasttwoimagesinwhichtheglaciersurfacetex- P
a
p (cid:74) (cid:73)
turecanbeobservedand(2)precisegeographiccoordinates,withanimageresolution e
r
20 superior to total glacier displacement during the observation period (Huang and Li, (cid:74) (cid:73)
|
2009). As long as the two conditions are satisfied, optical image cross-correlation can
becalculated(Ferrignoetal.,1994;Ertenetal.,2012).Thecontinentalglaciersonthe D Back Close
is
AntarcticPeninsulahaveasignificantlyhigheraveragevelocitythanmountainglaciers. cu FullScreen/Esc
s
Thus, medium-resolution MODIS data can meet the conditions for optical image fea- s
io
25 turetracking.Inaddition,MODISdatahaveadvantagesoffreeaccess,widecoverage, n Printer-friendlyVersion
P
short revisit cycles, and long time series. These advantages can help choose suitable a
p
e InteractiveDiscussion
r
5880 |
dataforcomparisonstudiesoverlongtimeseries,facilitatingcross-correlationcalcula- D
is
tion in large areas of the Antarctic region. c TCD
u
s
TheMODISL1BdataproductsareprovidedbyUSNationalAeronauticsandSpace s
io 8,5875–5910,2014
Administration(NASA).Inaccordancewiththetypeofsatellitecarrier,thedatacanbe n
P
5 dividedintotwocategories:MOD02(Terra-MODIS)andMYD02(Aqua-MODIS).Here, a
p
weuseband1data(spectralrange:620–670nm)oftheMOD02productat250mreso- e Spatiotemporal
r
lution.Band1L1Bdataareproductswithgeographiccoordinatesobtainedafterinstru- variations in the
|
mentcalibration.However,“geographicdata”and“scientificdata”thatrecordreflectivity surface velocities of
D
are not connected, resulting in so-called bow-tie phenomena near the edge when dis- Antarctic Peninsula
is
c
10 played directly. It is thus necessary to connect the MOD02 and MOD03 data products u glaciers
s
(Geolocation 1km) for geocoding prior to cross-correlation calculations. During cali- s
io
bration, the pixel values are synchronically converted from reflectance to DN values, n J.Chenetal.
P
ultimatelyproducingimageswithprecisegeographiccoordinates.BecausetheAntarc- a
p
e
tic Peninsula experiences cloudy, snowy weather and its winter months are strongly r
affected by the “polar night”, it is difficult to obtain high-quality continuous images. TitlePage
15 |
Therefore, we choose images from 2000, 2003, 2006, 2009, and 2012 data associ- Abstract Introduction
D
ated with little or no clouds for glacier velocity estimation (Fig. 2, Table 1). is
c Conclusions References
In addition, the Advanced Spaceborne Thermal Emission and Reflection Radiome- u
s
s
ter Global Digital Elevation Model (hereafter referred to as ASTER GDEM) within io Tables Figures
n
the study area is used, this version is developed in accordance with detailed Terra
20 P
a
observations (Holger and Frank, 2012), almost 1.3 million scenes of ASTER VNIR p (cid:74) (cid:73)
e
data covering the Earth land surface between 83◦N and 83◦S latitudes acquired from r
(cid:74) (cid:73)
1999 to 2010 have been processed by stereo-correlating (Slater et al., 2009). The |
ASTER GDEM was released in geotiff format with geographic lat/long coordinates, D Back Close
the spatial resolution is 30m and the vertical accuracy is 20m (Kang and Feng, is
25 c
u FullScreen/Esc
2011), its latitude–longitude coordinate was referenced to UTM/WGS 84 and the el- s
s
evations were computed with respect to the Earth Geopotential Model 1996 (EGM96) io
n
Printer-friendlyVersion
geoid (Ni et al., 2014), which meet the conditions for digital elevation calculations at P
a
larger spatial scales (ASTER GDEM data can be downloaded from the NASA web pe InteractiveDiscussion
r
5881 |
pages, http://asterweb.jpl.nasa.gov/gdem.asp or from the LPDAAC Global Data Ex- D
is
plorer) (Djamel and Hammadi, 2014). c TCD
u
s
s
io 8,5875–5910,2014
n
4 Methods P
a
p
e Spatiotemporal
The estimation of ice velocity using COSI-Corr mainly involves four steps: in-plane r
variations in the
5 displacement measurement, surface velocity calculation, three-dimensional (hereafter |
surface velocities of
referred to as 3-D) terrain correction, and the displacement orientation conversion
D
Antarctic Peninsula
(Fig. 3). is
c
u glaciers
s
s
4.1 Displacement measurement io
n J.Chenetal.
P
Early images are used as the reference images, and late images serve as the search a
p
e
10 images. Grid segmentation is performed to divide the reference images into standard r
TitlePage
windows(cells).Windows(cells)ofthesamesizethathavethehighestcorrelationare
|
then identified from the search area of the search image. The displacement between Abstract Introduction
D
thefeaturepointofmaximumcorrelationinthereferenceimageandthematchingpoint is
c Conclusions References
inthesearchimageisdeterminedaftergridsegmentationandthecross-correlationcal- u
s
s
15 culation.Aftercompletionofthecross-correlationcalculationinonegrid,itisshiftedby io Tables Figures
n
acertainstepsizeforcontinuouscalculationinthenextgrid.Thisprocedureeventually P
a
produces a raster data file with the step size as the pixel resolution and the matching p (cid:74) (cid:73)
e
featurepointsasthepixelvalues(Fig.4).ItshouldbenotedthatUTMprojection(same r
(cid:74) (cid:73)
length and width of each pixel) is adopted in the orthophoto coordinate system to fa- |
20 cilitategridsegmentation.However,duetoprojectiondeformationinthepolarregions, D Back Close
thereisanecessitytoconverttheresultsofthecross-correlationcalculationsintopolar is
c
azimuthalprojection(differentlengthandwidthofeachpixel)priortodisplacementand us FullScreen/Esc
s
area measurements, which will ensure the accuracy of the measurement. io
n
Printer-friendlyVersion
To improve the accuracy of correlation calculations, the grid size should be as large P
a
25 as possible and a larger number of pixels should be involved in the calculation. For pe InteractiveDiscussion
r
5882 |
example, the reference window for feature tracking of TM images in the Koxkar Baxi D
Glacier (average velocity <120ma−1) is set to 32×32 pixels (Xu et al., 2011). Addi- isc TCD
u
s
tionally,thesearchareashouldbeenlargedasmuchaspossibleandthespatialscale s
io 8,5875–5910,2014
of feature point matching can be extended to ensure that the possible maximum dis- n
P
5 placement of the glacier is included in the search area (Huang and Li, 2009). Haug a
p
et al. (2010) set the MODIS image-based search window to 44×44 pixels and the e Spatiotemporal
r
TMimage-basedsearchwindowto350×350pixelsfortheLarsenIceShelf.However, variations in the
|
theAntarcticPeninsulaglaciersurfacesaresohomogeneousthattherearefewdistin- surface velocities of
D
guishing feature points. In this case, excessively large grid and search areas are likely Antarctic Peninsula
is
c
10 to cause false matching of feature points. Based on a large amount of experimental u glaciers
s
analysis, we set the reference window for cross-correlation calculations to 10×10 pix- s
io
els, which corresponds to a ground distance of 2500m×2500m, and the search area n J.Chenetal.
P
is set to 36×36 pixels, which corresponds to a ground distance of 9000m×9000m. a
p
e
The cross-correlation calculation is performed after grid segmentation. Available al- r
TitlePage
gorithms include Orientation correlation (OC) (Haug et al., 2010), NCC (Kääb et al.,
15 |
2005), CCF (Heid and Kääb, 2012), and CCC (Huang and Li, 2011). In the present Abstract Introduction
D
study, we expect to identify the general feature and spatiotemporal variation patterns is
c Conclusions References
ofAntarcticPeninsulaglaciersurfacemovementthroughlarge-scale,long-durationve- u
s
s
locity estimation. Thus, there is no need for high-accuracy estimation of the velocity of io Tables Figures
n
thin ice streams within the glaciers. Furthermore, to reduce the amount of calculation,
20 P
a
wechosethesimpleCCCalgorithmthatprovidesmoderateaccuracyandinvolvesless p (cid:74) (cid:73)
e
data processing (Wu et al., 1997; Evans, 2000): r
(cid:74) (cid:73)
|
M N
(cid:80) (cid:80)[f(x,y)−f] D Back Close
CCC(u,v)= y=1x=1 · [g(x+u,y+v)−g(u,v)]2 (1) iscu FullScreen/Esc
s
(cid:40) (cid:41)1/2 (cid:40) (cid:41)1/2 s
(cid:80)M (cid:80)N [f(x,y)−f]2 (cid:80)M (cid:80)N [g(x+u,y+v)−g(u,v)]2 ion Printer-friendlyVersion
P
y=1x=1 y=1x=1 a
p
e InteractiveDiscussion
r
5883 |
where f(x,y) is the pixel value of point P in the reference window of the reference D
image; g(x,y) is the pixel value of point P(cid:48) in the corresponding search window of the isc TCD
u
searchimage(Fig.4);u,v arethedistancesofwindowmovementsinthesearchimage ss
io 8,5875–5910,2014
in the x axis and y axis directions, respectively; f,g(u,v) is the average pixel value of n
P
matching windows in the reference image and search image, respectively; N,M is the a
5 p
scope of the search area; and [f(x,y)−f] and [g(x+u,y+v)−g(u,v)] represent the er Spatiotemporal
difference between the DN value of each pixel and the average of all pixels within variations in the
|
a window, respectively. Thus, when a difference in the overall DN value of two images surface velocities of
D
fromtwosetsofMOD02dataindifferenttimephasesoccurs,e.g.duetovaryingofsun is Antarctic Peninsula
c
elevation angles, sensor angles or system error in atmospheric conditions, the CCC u glaciers
10 s
algorithmwillautomaticallyeliminatetheerrorinfluencethroughdifferentiationbetween sio
n J.Chenetal.
theindividualpixelvalueandtheaveragevalue.However,inthismethod,thedefinition P
a
ofafeaturepointistooabsolute,likelycausingfalsematchingoffeaturepoints(Huang p
e
and Li, 2009). Moreover, the Antarctic Peninsula has homogenous glacier surfaces r
TitlePage
15 with relatively smooth pixel values. Therefore, atmospheric and terrain corrections will |
inevitably screen out certain feature points, such as the texture in ice streams and the Abstract Introduction
D
ice mounds. Hence, we omit the atmospheric and terrain correction steps in the data is
c Conclusions References
u
processing. s
s
In the CCC algorithm, the results of each window calculation are recorded using io Tables Figures
n
the north–south and east–west displacement distances and the signal-to-noise ratio P
20 a
(hereafter referred to as SNR), and the results are then saved in a raster format. The p (cid:74) (cid:73)
e
r
pixel resolution of this raster file is the step size of the window movement (each step
(cid:74) (cid:73)
is a pixel, corresponding to a ground distance of 250m). The SNR directly records the |
credibilityofthisfeaturetrackingcalculation(Xuetal.,2011):SNR=1indicatesperfect D Back Close
correlation in the feature point matching, and SNR=0 indicates a perfect absence of isc
25 u FullScreen/Esc
correlation in the feature point matching. Thus, an SNR value close to 1 indicates high ss
io
credibility of the calculation results. We remove the results with SNR<0.7 to ensure n
Printer-friendlyVersion
P
the reliability of the correlation calculations. a
p
e InteractiveDiscussion
r
5884 |