1imageryintro(1) Grass User's Manual imageryintro(1)
2
6 Image processing in general
7 Digital numbers and physical values (reflection/radiance-at-sensor):
8 Satellite imagery is commonly stored in Digital Numbers (DN) for mini‐
10 mizing the storage volume, i.e. the originally sampled analog physical
11 value (color, temperature, etc) is stored a discrete representation in
12 8-16 bits. For example, Landsat data are stored in 8bit values (i.e.,
13 ranging from 0 to 255); other satellite data may be stored in 10 or 16
14 bits. Having data stored in DN, it implies that these data are not yet
15 the observed ground reality. Such data are called "at-satellite", for
16 example the amount of energy sensed by the sensor of the satellite
17 platform is encoded in 8 or more bits. This energy is called radi‐
18 ance-at-sensor. To obtain physical values from DNs, satellite image
19 providers use a linear transform equation (y = a * x + b) to encode the
20 radiance-at-sensor in 8 to 16 bits. DNs can be turned back into physi‐
21 cal values by applying the reverse formula (x = (y - b) / a).
22 The GRASS GIS module i.landsat.toar easily transforms Landsat DN to
24 radiance-at-sensor (top of atmosphere, TOA). The equivalent module for
25 ASTER data is i.aster.toar. For other satellites, r.mapcalc can be
26 employed.
27 Reflection/radiance-at-sensor and surface reflectance
29 When radiance-at-sensor has been obtained, still the atmosphere influ‐
31 ences the signal as recorded at the sensor. This atmospheric interac‐
32 tion with the sun energy reflected back into space by ground/vegeta‐
33 tion/soil needs to be corrected. The need of removing atmospheric arti‐
34 facts stems from the fact that the atmosphericic conditions are chang‐
35 ing over time. Hence, to gain comparability between Earth surface
36 images taken at different times, atmospheric need to be removed con‐
37 verting at-sensor values which are top of atmosphere to surface
38 reflectance values.
39 In GRASS GIS, there are two ways to apply atmospheric correction for
41 satellite imagery. A simple, less accurate way for Landsat is with
42 i.landsat.toar, using the DOS correction method. The more accurate way
43 is using i.atcorr (which supports many satellite sensors). The atmo‐
44 spherically corrected sensor data represent surface reflectance, which
45 ranges theoretically from 0% to 100%. Note that this level of data cor‐
46 rection is the proper level of correction to calculate vegetation
47 indices.
48 In GRASS GIS, image data are identical to raster data. However, a cou‐
50 ple of commands are explicitly dedicated to image processing. The geo‐
51 graphic boundaries of the raster/imagery file are described by the
52 north, south, east, and west fields. These values describe the lines
53 which bound the map at its edges. These lines do NOT pass through the
54 center of the grid cells at the edge of the map, but along the edge of
55 the map itself.
56 As a general rule in GRASS:
58 1 Raster/imagery output maps have their bounds and resolution
60 equal to those of the current region.
61 2 Raster/imagery input maps are automatically cropped/padded and
63 rescaled (using nearest-neighbor resampling) to match the cur‐
64 rent region.
65 Imagery import
67 The module r.in.gdal offers a common interface for many different
68 raster and satellite image formats. Additionally, it also offers
69 options such as on-the-fly location creation or extension of the
70 default region to match the extent of the imported raster map. For
71 special cases, other import modules are available. Always the full map
72 is imported. Imagery data can be group (e.g. channel-wise) with
73 i.group.
74 For importing scanned maps, the user will need to create a x,y-loca‐
76 tion, scan the map in the desired resolution and save it into an appro‐
77 priate raster format (e.g. tiff, jpeg, png, pbm) and then use r.in.gdal
78 to import it. Based on reference points the scanned map can be recti‐
79 fied to obtain geocoded data.
80 Image processing operations
82 GRASS raster/imagery map processing is always performed in the current
83 region settings (see g.region), i.e. the current region extent and cur‐
84 rent raster resolution is used. If the resolution differs from that of
85 the input raster map(s), on-the-fly resampling is performed (nearest
86 neighbor resampling). If this is not desired, the input map(s) has/have
87 to be resampled beforehand with one of the dedicated modules.
88 Geocoding of imagery data
90 GRASS is able to geocode raster and image data of various types:
91 · unreferenced scanned maps by defining four corner points
93 (i.group, i.target, g.gui.gcp, i.rectify)
94 · unreferenced satellite data from optical and Radar sensors by
96 defining a certain number of ground control points (i.group,
97 i.target, g.gui.gcp, i.rectify)
98 · interactive graphical Ground Control Point (GCP) manager
100 · orthophoto generation based on DEM: i.ortho.photo
102 · digital handheld camera geocoding: modified procedure for
104 i.ortho.photo
105 Visualizing (true) color composites
107 To quickly combine the first three channels to a near natural color
108 image, the GRASS command d.rgb can be used or the graphical GIS manager
109 (wxGUI). It assigns each channel to a color which is then mixed while
110 displayed. With a bit more work of tuning the grey scales of the chan‐
111 nels, nearly perfect colors can be achieved. Channel histograms can be
112 shown with d.histogram.
113 Calculation of vegetation indices
115 An example for indices derived from multispectral data is the NDVI
116 (normalized difference vegetation index). To study the vegetation sta‐
117 tus with NDVI, the Red and the Near Infrared channels (NIR) are taken
118 as used as input for simple map algebra in the GRASS command r.mapcalc
119 (ndvi = 1.0 * (nir - red)/(nir + red)). With r.colors an optimized
120 "ndvi" color table can be assigned afterward. Also other vegetation
121 indices can be generated likewise.
122 Calibration of thermal channel
124 The encoded digital numbers of a thermal infrared channel can be trans‐
125 formed to degree Celsius (or other temperature units) which represent
126 the temperature of the observed land surface. This requires a few alge‐
127 braic steps with r.mapcalc which are outlined in the literature to
128 apply gain and bias values from the image metadata.
129 Image classification
131 Single and multispectral data can be classified to user defined land
132 use/land cover classes. In case of a single channel, segmentation will
133 be used. GRASS supports the following methods:
134 · Radiometric classification:
136 · Unsupervised classification (i.cluster, i.maxlik) using the
138 Maximum Likelihood classification method
139 · Supervised classification (i.gensig or g.gui.iclass, i.max‐
141 lik) using the Maximum Likelihood classification method
142 · Combined radiometric/geometric (segmentation based) classifica‐
144 tion:
145 · Supervised classification (i.gensigset, i.smap)
147 · Object-oriented classification:
149 · Unsupervised classification (segmentation based: i.segment)
151 Kappa statistic can be calculated to validate the results (r.kappa).
152 Covariance/correlation matrices can be calculated with r.covar.
153 Image fusion
155 In case of using multispectral data, improvements of the resolution can
156 be gained by merging the panchromatic channel with color channels.
157 GRASS provides the HIS (i.rgb.his, i.his.rgb) and the Brovey and PCA
158 transform (i.pansharpen) methods.
159 Radiometric corrections
161 Atmospheric effects can be removed with i.atcorr. Correction for topo‐
162 graphic/terrain effects is offered in i.topo.corr. Clouds in LANDSAT
163 data can be identified and removed with i.landsat.acca. Calibrated
164 digital numbers of LANDSAT and ASTER imagery may be converted to
165 top-of-atmosphere radiance or reflectance and temperature
166 (i.aster.toar, i.landsat.toar).
167 Time series processing
169 GRASS also offers support for time series processing (r.series). Sta‐
170 tistics can be derived from a set of coregistered input maps such as
171 multitemporal satellite data. The common univariate statistics and also
172 linear regression can be calculated.
173 Evapotranspiration modeling
175 In GRASS, several types of evapotranspiration (ET) modeling methods are
176 available:
177 · Reference ET: Hargreaves (i.evapo.mh), Penman-Monteith
179 (i.evapo.pm);
180 · Potential ET: Priestley-Taylor (i.evapo.pt);
182 · Actual ET: i.evapo.time.
184 Evaporative fraction: i.eb.evapfr, i.eb.hsebal01.
185 Energy balance
187 Emissivity can be calculated with i.emissivity. Several modules sup‐
188 port the calculation of the energy balance:
189 · Actual evapotranspiration for diurnal period (i.eb.eta);
191 · Evaporative fraction and root zone soil moisture (i.eb.evapfr);
193 · Sensible heat flux iteration (i.eb.hsebal01);
195 · Net radiation approximation (i.eb.netrad);
197 · Soil heat flux approximation (i.eb.soilheatflux).
199 See also
201 · GRASS GIS Wiki page: Image processing
202 · The GRASS 4 Image Processing manual
204 · Introduction into raster data processing
206 · Introduction into 3D raster data (voxel) processing
208 · Introduction into vector data processing
210 · Introduction into temporal data processing
212 · Database management
214 · Projections and spatial transformations
216
218 Available at: Image processing in GRASS GIS source code (history)
219 Main index | Imagery index | Topics index | Keywords index | Graphical
221 index | Full index
222 © 2003-2019 GRASS Development Team, GRASS GIS 7.4.4 Reference Manual
224GRASS 7.4.4 imageryintro(1)