1r.geomorphon(1) GRASS GIS User's Manual r.geomorphon(1)
2
6 r.geomorphon - Calculates geomorphons (terrain forms) and associated
7 geometry using machine vision approach.
8
10 raster, geomorphons, terrain patterns, machine vision geomorphometry
11
13 r.geomorphon
14 r.geomorphon --help
15 r.geomorphon [-me] elevation=name [forms=name] [ternary=name]
16 [positive=name] [negative=name] [intensity=name] [exposi‐
17 tion=name] [range=name] [variance=name] [elongation=name] [az‐
18 imuth=name] [extend=name] [width=name] search=integer skip=integer
19 flat=float dist=float [comparison=string] [coordinates=east,north]
20 [profiledata=name] [profileformat=string] [--overwrite] [--help]
21 [--verbose] [--quiet] [--ui]
22 Flags:
24 -m
25 Use meters to define search units (default is cells)
26 -e
28 Use extended form correction
29 --overwrite
31 Allow output files to overwrite existing files
32 --help
34 Print usage summary
35 --verbose
37 Verbose module output
38 --quiet
40 Quiet module output
41 --ui
43 Force launching GUI dialog
44 Parameters:
46 elevation=name [required]
47 Name of input elevation raster map
48 forms=name
50 Most common geomorphic forms
51 ternary=name
53 Code of ternary patterns
54 positive=name
56 Code of binary positive patterns
57 negative=name
59 Code of binary negative patterns
60 intensity=name
62 Rasters containing mean relative elevation of the form
63 exposition=name
65 Rasters containing maximum difference between extend and central
66 cell
67 range=name
69 Rasters containing difference between max and min elevation of the
70 form extend
71 variance=name
73 Rasters containing variance of form boundary
74 elongation=name
76 Rasters containing local elongation
77 azimuth=name
79 Rasters containing local azimuth of the elongation
80 extend=name
82 Rasters containing local extend (area) of the form
83 width=name
85 Rasters containing local width of the form
86 search=integer [required]
88 Outer search radius
89 Default: 3
90 skip=integer [required]
92 Inner search radius
93 Default: 0
94 flat=float [required]
96 Flatness threshold (degrees)
97 Default: 1
98 dist=float [required]
100 Flatness distance, zero for none
101 Default: 0
102 comparison=string
104 Comparison mode for zenith/nadir line-of-sight search
105 Options: anglev1, anglev2, anglev2_distance
106 Default: anglev1
107 coordinates=east,north
109 Coordinates to profile
110 profiledata=name
112 Profile output file name ("-" for stdout)
113 profileformat=string
115 Profile output format
116 Options: json, yaml, xml
117
119 What is geomorphon:
120 Geomorphon is a new concept of presentation and analysis of terrain
121 forms. This concept utilises 8-tuple pattern of the visibility neigh‐
122 bourhood and breaks well known limitation of standard calculus ap‐
123 proach where all terrain forms cannot be detected in a single window
124 size. The pattern arises from a comparison of a focus pixel with its
125 eight neighbors starting from the one located to the east and continu‐
126 ing counterclockwise producing ternary operator. For example, a tuple
127 {+,-,-,-,0,+,+,+} describes one possible pattern of relative measures
128 {higher, lower, lower, lower, equal, higher, higher, higher} for pixels
129 surrounding the focus pixel. It is important to stress that the visi‐
130 bility neighbors are not necessarily an immediate neighbors of the fo‐
131 cus pixel in the grid, but the pixels determined from the line-of-sight
132 principle along the eight principal directions. This principle relates
133 surface relief and horizontal distance by means of so-called zenith and
134 nadir angles along the eight principal compass directions. The ternary
135 operator converts the information contained in all the pairs of zenith
136 and nadir angles into the ternary pattern (8-tuple). The result depends
137 on the values of two parameters: search radius (L) and relief threshold
138 (d). The search radius is the maximum allowable distance for calcula‐
139 tion of zenith and nadir angles. The relief threshold is a minimum
140 value of difference between LOSs angle (zenith and nadir) that is con‐
141 sidered significantly different from the horizon. Two lines-of-sight
142 are necessary due to zenith LOS only, does not detect positive forms
143 correctly.
144 There are 3**8 = 6561 possible ternary patterns (8-tuples). However by
146 eliminating all patterns that are results of either rotation or reflec‐
147 tion of other patterns wa set of 498 patterns remain referred as geo‐
148 morphons. This is a comprehensive and exhaustive set of idealized
149 landforms that are independent of the size, relief, and orientation of
150 the actual landform.
151 Form recognition depends on two free parameters: Search radius and
153 flatness threshold. Using larger values of L and is tantamount to ter‐
154 rain classification from a higher and wider perspective, whereas using
155 smaller values of L and is tantamount to terrain classification from a
156 local point of view. A character of the map depends on the value of L.
157 Using small value of L results in the map that correctly identifies
158 landforms if their size is smaller than L; landforms having larger
159 sizes are broken down into components. Using larger values of L allows
160 simultaneous identification of landforms on variety of sizes in expense
161 of recognition smaller, second-order forms. There are two additional
162 parameters: skip radius used to eliminate impact of small irregulari‐
163 ties. On the contrary flatness distance eliminates the impact of very
164 high distance (in meters) of search radius which may not detect eleva‐
165 tion difference if this is at very far distance. Important especially
166 with low resolution DEMS.
167
169 -m
170 All distance parameters (search, skip, flat distances) are supplied
171 as meters instead of cells (default). To avoid situation when sup‐
172 plied distances is smaller than one cell program first check if
173 supplied distance is longer than one cell in both NS and WE direc‐
174 tions. For LatLong projection only NS distance checked, because in
175 latitude angular unit comprise always bigger or equal distance than
176 longitude one. If distance is supplied in cells, For all projec‐
177 tions is recalculated into meters according formula: num‐
178 ber_of_cells*resolution_along_NS_direction. It is important if geo‐
179 morphons are calculated for large areas in LatLong projection.
180 elevation
182 Digital elevation model. Data can be of any type and any projec‐
183 tion. During calculation DEM is stored as floating point raster.
184 search
186 Determines length on the geodesic distances in all eight directions
187 where line-of-sight is calculated. To speed up calculation is de‐
188 termines only these cells which centers falls into the distance.
189 skip
191 Determines length on the geodesic distances at the beginning of
192 calculation all eight directions where line-of-sight is yet calcu‐
193 lated. To speed up calculation this distance is always recalculated
194 into number of cell which are skipped at the beginning of every
195 line-of-sight and is equal in all direction. This parameter elimi‐
196 nates forms of very small extend, smaller than skip parameter.
197 flat
199 The difference (in degrees) between zenith and nadir line-of-sight
200 which indicate flat direction. If higher threshold produce more
201 flat maps. If resolution of the map is low (more than 1 km per
202 cell) threshold should be very small (much smaller than 1 degree)
203 because on such distance 1 degree of difference means several me‐
204 ters of high difference.
205 dist
207 >Flat distance. This is additional parameter defining the distance
208 above which the threshold starts to decrease to avoid problems with
209 pseudo-flat line-of-sights if real elevation difference appears on
210 the distance where its value is higher (TO BE CORRECTED).
211 comparison
213 Comparison mode for zenith/nadir line-of-sight search. "anglev1" is
214 the original r.geomorphon comparison mode. "anglev2" is an improved
215 mode, which better handles angle thresholds and zenith/nadir angles
216 that are exactly equal. "anglev2_distance" in addition to that
217 takes the zenith/nadir distances into account when the angles are
218 exactly equal.
219 forms
221 Returns geomorphic map with 10 most popular terrestrial forms. Leg‐
222 end for forms, its definition by the number of + and - and its ide‐
223 alized visualisation are presented at the image.
224 Forms represented by geomorphons:
226 ternary
227 returns code of one of 498 unique ternary patterns for every cell.
228 The code is a decimal representation of 8-tuple minimalised pat‐
229 terns written in ternary system. Full list of patterns is available
230 in source code directory as patterns.txt. This map can be used to
231 create alternative form classification using supervised approach.
232 positive and negative
234 returns codes binary patterns for zenith (positive) and nadir (neg‐
235 ative) line of sights. The code is a decimal representation of
236 8-tuple minimalised patterns written in binary system. Full list of
237 patterns is available in source code directory as patterns.txt.
238 coordinates
240 The central point of a single geomorphon to profile. The central
241 point must be within the computational region, which should be
242 large enough to accommodate the search radius. Setting the region
243 larger than that will not produce more accurate data, but in the
244 current implementation will slow the computation down. For the best
245 results remember to align the region to the raster cells. Profiling
246 is mutually exclusive with any raster outputs, but other parameters
247 and flags (such as elevation, search, comparison, -m and -e) work
248 as usual.
249 profiledata
251 The output file name for the complete profile data, "-" means to
252 write to the standard output. The data is in a machine-readable
253 format and it includes assorted values describing the computation
254 context and parameters, as well as its intermediate and final re‐
255 sults.
256 profileformat
258 Format of the profile data: "json", "yaml" or "xml".
259 NOTE: parameters below are experimental. The usefulness of these param‐
261 eters are currently under investigation.
262 intensity
264 returns avarage difference between central cell of geomorphon and
265 eight cells in visibility neighbourhood. This parameter shows local
266 (as is visible) exposition/abasement of the form in the terrain.
267 range
269 returns difference between minimum and maximum values of visibility
270 neighbourhood.
271 variance
273 returns variance (difference between particular values and mean
274 value) of visibility neighbourhood.
275 extend
277 returns area of the polygon created by the 8 points where
278 line-of-sight cuts the terrain (see image in description section).
279 azimuth
281 returns orientation of the polygon constituting geomorphon. This
282 orientation is currently calculated as a orientation of least
283 square fit line to the eight verticles of this polygon.
284 elongation
286 returns proportion between sides of the bounding box rectangle cal‐
287 culated for geomorphon rotated to fit least square line.
288 width
290 returns length of the shorter side of the bounding box rectangle
291 calculated for geomorphon rotated to fit least square line.
292
294 From computational point of view there are no limitations of input DEM
295 and free parameters used in calculation. However, in practice there are
296 some issues on DEM resolution and search radius. Low resolution DEM es‐
297 pecially above 1 km per cell requires smaller than default flatness
298 threshold. On the other hand, only forms with high local elevation dif‐
299 ference will be detected correctly. It results from fact that on very
300 high distance (of order of kilometers or higher) even relatively high
301 elevation difference will be recognized as flat. For example at the
302 distance of 8 km (8 cells with 1 km resolution DEM) an relative eleva‐
303 tion difference of at least 136 m is required to be noticed as
304 non-flat. Flatness distance threshold may be helpful to avoid this
305 problem.
306
308 Geomorphon calculation: extraction of terrestrial landforms
309 Geomorphon calculation example using the EU DEM 25m:
310 g.region raster=eu_dem_25m -p
311 r.geomorphon elevation=eu_dem_25m forms=eu_dem_25m_geomorph
312 # verify terrestrial landforms found in DEM
313 r.category eu_dem_25m_geomorph
314 1 flat
315 2 peak
316 3 ridge
317 4 shoulder
318 5 spur
319 6 slope
320 7 hollow
321 8 footslope
322 9 valley
323 10 pit
324 Extraction of peaks
326 Using the resulting terrestrial landforms map, single landforms can be
327 extracted, e.g. the peaks, and converted into a vector point map:
328 r.mapcalc expression="eu_dem_25m_peaks = if(eu_dem_25m_geomorph == 2, 1, null())"
329 r.thin input=eu_dem_25m_peaks output=eu_dem_25m_peaks_thinned
330 r.to.vect input=eu_dem_25m_peaks_thinned output=eu_dem_25m_peaks type=point
331 v.info input=eu_dem_25m_peaks
332
334 • Stepinski, T., Jasiewicz, J., 2011, Geomorphons - a new ap‐
335 proach to classification of landform, in : Eds: Hengl, T.,
336 Evans, I.S., Wilson, J.P., and Gould, M., Proceedings of Geo‐
337 morphometry 2011, Redlands, 109-112 (PDF)
338 • Jasiewicz, J., Stepinski, T., 2013, Geomorphons - a pattern
340 recognition approach to classification and mapping of land‐
341 forms, Geomorphology, vol. 182, 147-156 (DOI: 10.1016/j.geo‐
342 morph.2012.11.005)
343
345 r.param.scale
346
348 Jarek Jasiewicz, Tomek Stepinski (merit contribution)
349
351 Available at: r.geomorphon source code (history)
352 Accessed: Mon Jun 20 16:46:04 2022
354 Main index | Raster index | Topics index | Keywords index | Graphical
356 index | Full index
357 © 2003-2022 GRASS Development Team, GRASS GIS 8.2.0 Reference Manual
359GRASS 8.2.0 r.geomorphon(1)