-
Notifications
You must be signed in to change notification settings - Fork 448
/
distancexform.m
316 lines (280 loc) · 10.7 KB
/
distancexform.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
%DISTANCEXFORM Distance transform
%
% D = DISTANCEXFORM(IM, OPTIONS) is the distance transform of the binary
% image IM. The elements of D have a value equal to the shortest distance
% from that element to a non-zero pixel in the input image IM.
%
% D = DISTANCEXFORM(OCCGRID, GOAL, OPTIONS) is the distance transform of
% the occupancy grid OCCGRID with respect to the specified goal point GOAL
% = [X,Y]. The cells of the grid have values of 0 for free space and 1 for
% obstacle. The resulting matrix D has cells whose value is the shortest
% distance to the goal from that cell, or NaN if the cell corresponds to an
% obstacle (set to 1 in OCCGRID).
%
% Options:
% 'euclidean' Use Euclidean (L2) distance metric (default)
% 'cityblock' Use cityblock or Manhattan (L1) distance metric
%
% 'animate' Show the iterations of the computation
% 'delay',D Delay of D seconds between animation frames (default 0.2s)
% 'movie',M Save animation to a movie file or folder
%
% 'noipt' Don't use Image Processing Toolbox, even if available
% 'novlfeat' Don't use VLFeat, even if available
% 'nofast' Don't use IPT, VLFeat or imorph, even if available.
%
% 'delay'
%
% Notes::
% - For the first case Image Processing Toolbox (IPT) or VLFeat will be used if
% available, searched for in that order. They use a 2-pass rather than
% iterative algorithm and are much faster.
% - Options can be used to disable use of IPT or VLFeat.
% - If IPT or VLFeat are not available, or disabled, then imorph is used.
% - If IPT, VLFeat or imorph are not available a slower M-function is used.
% - If the 'animate' option is given then the MATLAB implementation is used.
% - Using imorph requires iteration and is slow.
% - For the second case the Machine Vision Toolbox function imorph is required.
% - imorph is a mex file and must be compiled.
% - The goal is given as [X,Y] not MATLAB [row,col] format.
%
% See also IMORPH, DXform, Animate.
% Copyright (C) 1993-2017, by Peter I. Corke
%
% This file is part of The Robotics Toolbox for MATLAB (RTB).
%
% RTB is free software: you can redistribute it and/or modify
% it under the terms of the GNU Lesser General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% RTB is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
% GNU Lesser General Public License for more details.
%
% You should have received a copy of the GNU Leser General Public License
% along with RTB. If not, see <http://www.gnu.org/licenses/>.
%
% http://www.petercorke.com
function dx = distancexform(occgrid, varargin)
opt.delay = 0.2;
opt.ipt = true;
opt.vlfeat = true;
opt.fast = true;
opt.metric = {'euclidean', 'cityblock'};
opt.animate = false;
opt.movie = [];
[opt,args] = tb_optparse(opt, varargin);
if opt.movie
opt.animate = true;
end
if opt.animate
opt.fast = false;
opt.ipt = false;
opt.vlfeat = false;
clf
end
count = [];
switch opt.metric
case 'cityblock'
ipt_metric = opt.metric; % if we use bwdistgeodesic
m = [ inf 1 inf
1 0 1
inf 1 inf ];
case 'euclidean'
ipt_metric = 'quasi-euclidean'; % if we use bwdistgeodesic
r2 = sqrt(2);
m = [ r2 1 r2
1 0 1
r2 1 r2 ];
end
if ~isempty(args) && isvec(args{1}, 2)
%% path planning interpretation
% distancexform(world, goal, metric, show)
goal = args{1};
occgrid = double(occgrid);
% check the goal point is sane
assert(occgrid(goal(2), goal(1)) == 0, 'RTB:distancexform:badarg', 'goal inside obstacle')
if exist('imorph', 'file') == 3 && opt.fast
% using MEX version of imorph
if opt.verbose
fprintf('using MVTB:imorph\n');
end
% setup to use imorph
% - set obstacles to NaN
% - set free space to Inf
% - set goal to 0
occgrid(occgrid>0) = NaN;
occgrid(occgrid==0) = Inf;
occgrid(goal(2), goal(1)) = 0;
count = 0;
ninf = Inf; % number of infinities in the map
while 1
occgrid = imorph(occgrid, m, 'plusmin');
count = count+1;
if opt.animate
cmap = [1 0 0; gray(count)];
colormap(cmap)
image(occgrid+1, 'CDataMapping', 'direct');
set(gca, 'Ydir', 'normal');
xlabel('x');
ylabel('y');
pause(opt.delay);
end
ninfnow = sum( isinf(occgrid(:)) ); % current number of Infs
if ninfnow == ninf
% stop if the number of Infs left in the map had stopped reducing
% it may never get to zero if there are unreachable cells in the map
break;
end
ninf = ninfnow;
end
dx = occgrid;
elseif exist('bwdistgeodesic', 'file') && opt.ipt
if opt.verbose
fprintf('using IPT:bwdistgeodesic\n');
end
% solve using IPT
dx = double( bwdistgeodesic(occgrid==0, goal(1), goal(2), ipt_metric) );
else
if opt.verbose
fprintf('using MATLAB code, faster if you install MVTB\n');
end
% setup to use M-function
occgrid(occgrid>0) = NaN;
nans = isnan(occgrid);
occgrid(occgrid==0) = Inf;
occgrid(goal(2), goal(1)) = 0;
count = 0;
ninf = Inf; % number of infinities in the map
anim = Animate(opt.movie);
while 1
occgrid = dxstep(occgrid, m);
occgrid(nans) = NaN;
count = count+1;
if opt.animate
cmap = [1 0 0; gray(count)];
colormap(cmap)
image(occgrid+1, 'CDataMapping', 'direct');
set(gca, 'Ydir', 'normal');
xlabel('x');
ylabel('y');
if opt.animate
anim.add();
else
pause(opt.delay);
end
end
ninfnow = sum( isinf(occgrid(:)) ); % current number of Infs
if ninfnow == ninf
% stop if the number of Infs left in the map had stopped reducing
% it may never get to zero if there are unreachable cells in the map
break;
end
ninf = ninfnow;
end
anim.close();
dx = occgrid;
end
if opt.animate && ~isempty(count)
fprintf('%d iterations, %d unreachable cells\n', count, ninf);
end
else
%% image processing interpretation
% distancexform(world, [metric])
if exist('imorph', 'file') && opt.fast
if opt.verbose
fprintf('using MVTB:imorph\n');
end
% setup to use imorph
% - set free space to Inf
% - set goal to 0
occgrid = double(occgrid);
occgrid(occgrid==0) = Inf;
occgrid(isfinite(occgrid)) = 0;
count = 0;
anim = Animate(opt.movie);
while 1
occgrid = imorph(occgrid, m, 'plusmin');
count = count+1;
if opt.show
cmap = [1 0 0; gray(count)];
colormap(cmap)
image(occgrid+1, 'CDataMapping', 'direct');
set(gca, 'Ydir', 'normal');
xlabel('x');
ylabel('y');
if opt.animate
anim.add();
else
pause(opt.delay);
end
end
ninfnow = sum( isinf(occgrid(:)) ); % current number of Infs
if ninfnow == 0
% stop if no Infs left in the image
break;
end
end
anim.close();
dx = occgrid;
elseif exist('bwdist') && opt.ipt
if opt.verbose
fprintf('using IPT:bwdist\n');
end
% use IPT
dx = bwdist(occgrid, ipt_metric);
elseif exist('vl_imdisttf') && opt.vlfeat
if opt.verbose
fprintf('using VLFEAT:vl_imsdisttf\n');
end
im = double(occgrid);
im(im==0) = inf;
im(im==1) = 0;
d2 = vl_imdisttf(im);
dx = sqrt(d2);
else
if opt.verbose
fprintf('using MATLAB code, faster if you install MVTB\n');
end
occgrid = double(occgrid);
occgrid(occgrid==0) = Inf;
occgrid(isfinite(occgrid)) = 0;
count = 0;
while 1
occgrid = dxstep(occgrid, m);
count = count+1;
if opt.show
cmap = [1 0 0; gray(count)];
colormap(cmap)
image(occgrid+1, 'CDataMapping', 'direct');
set(gca, 'Ydir', 'normal');
xlabel('x');
ylabel('y');
pause(opt.show);
end
ninfnow = sum( isinf(occgrid(:)) ); % current number of Infs
if ninfnow == 0
% stop if no Infs left in the image
break;
end
end
dx = occgrid;
end
end
end
% MATLAB implementation of computational kernel
function out = dxstep(G, m)
[h,w] = size(G); % get size of occ grid
% pad with inf
G = [ones(1,w)*Inf; G; ones(1,w)*Inf];
G = [ones(h+2,1)*Inf G ones(h+2,1)*Inf];
w = w+2; h = h+2;
for r=2:h-1
for c=2:w-1
W = G(r-1:r+1,c-1:c+1); % get 3x3 window
out(r-1,c-1) = min(min(W+m)); % add distances and find the minimum
end
end
end