1v.lidar.edgedetection(1) Grass User's Manual v.lidar.edgedetection(1)
2
6 v.lidar.edgedetection - Detects the object’s edges from a LIDAR data
7 set.
8
10 vector, LIDAR, edges
11
13 v.lidar.edgedetection
14 v.lidar.edgedetection --help
15 v.lidar.edgedetection [-e] input=name output=name [ew_step=float]
16 [ns_step=float] [lambda_g=float] [tgh=float] [tgl=float]
17 [theta_g=float] [lambda_r=float] [--overwrite] [--help] [--ver‐
18 bose] [--quiet] [--ui]
19 Flags:
21 -e
22 Estimate point density and distance and quit
23 Estimate point density and distance in map units for the input vec‐
24 tor points within the current region extents and quit
25 --overwrite
27 Allow output files to overwrite existing files
28 --help
30 Print usage summary
31 --verbose
33 Verbose module output
34 --quiet
36 Quiet module output
37 --ui
39 Force launching GUI dialog
40 Parameters:
42 input=name [required]
43 Name of input vector map
44 Or data source for direct OGR access
45 output=name [required]
47 Name for output vector map
48 ew_step=float
50 Length of each spline step in the east-west direction
51 Default: 4 * east-west resolution
52 ns_step=float
54 Length of each spline step in the north-south direction
55 Default: 4 * north-south resolution
56 lambda_g=float
58 Regularization weight in gradient evaluation
59 Default: 0.01
60 tgh=float
62 High gradient threshold for edge classification
63 Default: 6
64 tgl=float
66 Low gradient threshold for edge classification
67 Default: 3
68 theta_g=float
70 Angle range for same direction detection
71 Default: 0.26
72 lambda_r=float
74 Regularization weight in residual evaluation
75 Default: 2
76
78 v.lidar.edgedetection is the first of three steps to filter LiDAR data.
79 The filter aims to recognize and extract attached and detached object
80 (such as buildings, bridges, power lines, trees, etc.) in order to
81 create a Digital Terrain Model.
82 In particular, this module detects the edge of each single feature over
84 the terrain surface of a LIDAR point surface. First of all, a bilinear
85 spline interpolation with a Tychonov regularization parameter is per‐
86 formed. The gradient is minimized and the low Tychonov regularization
87 parameter brings the interpolated functions as close as possible to the
88 observations. Bicubic spline interpolation with Tychonov regularization
89 is then performed. However, now the curvature is minimized and the reg‐
90 ularization parameter is set to a high value. For each point, an inter‐
91 polated value is computed from the bicubic surface and an interpolated
92 gradient is computed from the bilinear surface. At each point the gra‐
93 dient magnitude and the direction of the edge vector are calculated,
94 and the residual between interpolated and observed values is computed.
95 Two thresholds are defined on the gradient, a high threshold tgh and a
96 low one tgl. For each point, if the gradient magnitude is greater than
97 or equal to the high threshold and its residual is greater than or
98 equal to zero, it is labeled as an EDGE point. Similarly a point is
99 labeled as being an EDGE point if the gradient magnitude is greater
100 than or equal to the low threshold, its residual is greater than or
101 equal to zero, and the gradient to two of eight neighboring points is
102 greater than the high threshold. Other points are classified as TER‐
103 RAIN.
104 The length (in mapping units) of each spline step is defined by ew_step
106 for the east-west direction and ns_step for the north-south direction.
107 The output will be a vector map in which points has been classified as
109 TERRAIN, EDGE or UNKNOWN. This vector map should be the input of
110 v.lidar.growing module.
111
113 In this module, an external table will be created which will be useful
114 for the next module of the procedure of LiDAR data filtering. In this
115 table the interpolated height values of each point will be recorded.
116 Also points in the output vector map will be classified as:
117 TERRAIN (cat = 1, layer = 1)
118 EDGE (cat = 2, layer = 1)
119 UNKNOWN (cat = 3, layer = 1)
120 The final result of the whole procedure (v.lidar.edgedetection,
121 v.lidar.growing, v.lidar.correction) will be a point classification in
122 four categories:
123 TERRAIN SINGLE PULSE (cat = 1, layer = 2)
124 TERRAIN DOUBLE PULSE (cat = 2, layer = 2)
125 OBJECT SINGLE PULSE (cat = 3, layer = 2)
126 OBJECT DOUBLE PULSE (cat = 4, layer = 2)
127
129 Basic edge detection
130 # last return points
131 v.lidar.edgedetection input=vector_last output=edge ew_step=8 ns_step=8 lambda_g=0.5
132 Complete workflow
134 # region settings (using an existing raster)
135 g.region raster=elev_lid792_1m
136 # import
137 v.in.lidar -tr input=points.las output=points
138 v.in.lidar -tr input=points.las output=points_first return_filter=first
139 # detection
140 v.lidar.edgedetection input=points output=edge ew_step=8 ns_step=8 lambda_g=0.5
141 v.lidar.growing input=edge output=growing first=points_first
142 v.lidar.correction input=growing output=correction terrain=only_terrain
143 # visualization of selected points
144 # zoom somewhere first, to make it faster
145 d.rast map=orthophoto
146 d.vect map=correction layer=2 cats=2,3,4 color=red size=0.25
147 d.vect map=correction layer=2 cats=1 color=0:128:0 size=0.5
148 # interpolation (this may take some time)
149 v.surf.rst input=only_terrain elevation=terrain
150 # get object points for 3D visualization
151 v.extract input=correction layer=2 cats=2,3,4 output=objects
152 Figure 1: Example output from complete workflow (red: objects, green:
154 terrain)
155 Figure 2: 3D visualization of filtered object points (red) and terrain
157 created from terrain points (gray)
158
160 · Antolin, R. et al., 2006. Digital terrain models determination
161 by LiDAR technology: Po basin experimentation. Bolletino di
162 Geodesia e Scienze Affini, anno LXV, n. 2, pp. 69-89.
163 · Brovelli M. A., Cannata M., Longoni U.M., 2004. LIDAR Data Fil‐
165 tering and DTM Interpolation Within GRASS, Transactions in GIS,
166 April 2004, vol. 8, iss. 2, pp. 155-174(20), Blackwell Pub‐
167 lishing Ltd.
168 · Brovelli M. A., Cannata M., 2004. Digital Terrain model recon‐
170 struction in urban areas from airborne laser scanning data: the
171 method and an example for Pavia (Northern Italy). Computers
172 and Geosciences 30 (2004) pp.325-331
173 · Brovelli M. A. and Longoni U.M., 2003. Software per il filtrag‐
175 gio di dati LIDAR, Rivista dell’Agenzia del Territorio, n.
176 3-2003, pp. 11-22 (ISSN 1593-2192).
177 · Brovelli M. A., Cannata M. and Longoni U.M., 2002. DTM LIDAR in
179 area urbana, Bollettino SIFET N.2, pp. 7-26.
180 · Performances of the filter can be seen in the ISPRS WG III/3
182 Comparison of Filters report by Sithole, G. and Vosselman, G.,
183 2003.
184
186 v.lidar.growing, v.lidar.correction, v.surf.bspline, v.surf.rst,
187 v.in.lidar, v.in.ascii
188
190 Original version of program in GRASS 5.4:
191 Maria Antonia Brovelli, Massimiliano Cannata, Ulisse Longoni and Mirko
192 Reguzzoni
193 Update for GRASS 6.X:
194 Roberto Antolin and Gonzalo Moreno
195
197 Available at: v.lidar.edgedetection source code (history)
198 Main index | Vector index | Topics index | Keywords index | Graphical
200 index | Full index
201 © 2003-2019 GRASS Development Team, GRASS GIS 7.8.2 Reference Manual
203GRASS 7.8.2 v.lidar.edgedetection(1)