1 | % * This code was used in the following articles:
|
---|
2 | % * [1] Learning 3-D Scene Structure from a Single Still Image,
|
---|
3 | % * Ashutosh Saxena, Min Sun, Andrew Y. Ng,
|
---|
4 | % * In ICCV workshop on 3D Representation for Recognition (3dRR-07), 2007.
|
---|
5 | % * (best paper)
|
---|
6 | % * [2] 3-D Reconstruction from Sparse Views using Monocular Vision,
|
---|
7 | % * Ashutosh Saxena, Min Sun, Andrew Y. Ng,
|
---|
8 | % * In ICCV workshop on Virtual Representations and Modeling
|
---|
9 | % * of Large-scale environments (VRML), 2007.
|
---|
10 | % * [3] 3-D Depth Reconstruction from a Single Still Image,
|
---|
11 | % * Ashutosh Saxena, Sung H. Chung, Andrew Y. Ng.
|
---|
12 | % * International Journal of Computer Vision (IJCV), Aug 2007.
|
---|
13 | % * [6] Learning Depth from Single Monocular Images,
|
---|
14 | % * Ashutosh Saxena, Sung H. Chung, Andrew Y. Ng.
|
---|
15 | % * In Neural Information Processing Systems (NIPS) 18, 2005.
|
---|
16 | % *
|
---|
17 | % * These articles are available at:
|
---|
18 | % * http://make3d.stanford.edu/publications
|
---|
19 | % *
|
---|
20 | % * We request that you cite the papers [1], [3] and [6] in any of
|
---|
21 | % * your reports that uses this code.
|
---|
22 | % * Further, if you use the code in image3dstiching/ (multiple image version),
|
---|
23 | % * then please cite [2].
|
---|
24 | % *
|
---|
25 | % * If you use the code in third_party/, then PLEASE CITE and follow the
|
---|
26 | % * LICENSE OF THE CORRESPONDING THIRD PARTY CODE.
|
---|
27 | % *
|
---|
28 | % * Finally, this code is for non-commercial use only. For further
|
---|
29 | % * information and to obtain a copy of the license, see
|
---|
30 | % *
|
---|
31 | % * http://make3d.stanford.edu/publications/code
|
---|
32 | % *
|
---|
33 | % * Also, the software distributed under the License is distributed on an
|
---|
34 | % * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
|
---|
35 | % * express or implied. See the License for the specific language governing
|
---|
36 | % * permissions and limitations under the License.
|
---|
37 | % *
|
---|
38 | % */
|
---|
39 | function [F, newinliers, fail] = RansacNonlinearReFinement(f1, f2, matches, F, inliers, I1, I2, disp) |
---|
40 | % first re-estimates F based on a non-linear method, then projects each |
---|
41 | % point into 3D space and removes matches that are either very close or |
---|
42 | % very far away from camera 1 |
---|
43 | |
---|
44 | numMatches = size(matches, 2); |
---|
45 | numInliers = size(inliers, 2); |
---|
46 | percentInliers = numInliers/numMatches; |
---|
47 | |
---|
48 | fail = 0; |
---|
49 | if (percentInliers < 0.2) | (numInliers < 20) |
---|
50 | fail = 1; |
---|
51 | newinliers = inliers; |
---|
52 | else |
---|
53 | x1 = [f1(:, matches(1, inliers))]; |
---|
54 | x2 = [f2(:, matches(2, inliers))]; |
---|
55 | x = [x1' x2']; |
---|
56 | m3=1; %256 |
---|
57 | |
---|
58 | % re-estimate the fundamental matrix using a non-linear method |
---|
59 | [f, f_sq_errors, n_inliers,inlier_index,F] = torr_estimateF( x, m3, [], 'non_linear', 0, F); |
---|
60 | |
---|
61 | % find the 3D positions of the points |
---|
62 | X = ThreeDPoints(F, size(I1), size(I2), ... |
---|
63 | f1(:, matches(1, inliers)), f2(:, matches(2, inliers)));%F'?? |
---|
64 | |
---|
65 | size(X) |
---|
66 | |
---|
67 | % find the distance to camera1, sort them, find the size of gaps |
---|
68 | % between points |
---|
69 | dist = sqrt(sum(X.^2)); |
---|
70 | sdist = sort(dist); |
---|
71 | spans = sdist - [0 sdist(1:end-1)]; |
---|
72 | |
---|
73 | % find the start/end dist of the middle 50% of points |
---|
74 | numpts = length(sdist) |
---|
75 | startLind = ceil(numpts*0.25); |
---|
76 | startHind = ceil(numpts*0.75); |
---|
77 | startL = sdist(startLind); |
---|
78 | startH = sdist(startHind); |
---|
79 | |
---|
80 | % don't allow spans greater than 20 times the average span of the |
---|
81 | % middle 50% of points |
---|
82 | maxSpan = 20*(startH - startL)/numpts; |
---|
83 | badspansH = find(spans>maxSpan); |
---|
84 | badspansL = find(spans>maxSpan/2); |
---|
85 | [lowidx, temp] = max(badspansL(badspansL<startLind)); |
---|
86 | if isempty(lowidx) |
---|
87 | lowidx=1; |
---|
88 | end |
---|
89 | [highidx, temp] = min(badspansH(badspansH>startHind)); |
---|
90 | if isempty(highidx) |
---|
91 | highidx=length(sdist); |
---|
92 | end |
---|
93 | lowbound = sdist(lowidx) |
---|
94 | highbound = sdist(highidx) |
---|
95 | keep = (dist>lowbound).*(dist<highbound); |
---|
96 | idx = find(keep); |
---|
97 | rejectIdx=find(1-keep); |
---|
98 | |
---|
99 | % get the new inliers |
---|
100 | newinliers = inliers(idx); |
---|
101 | numNewInliers = length(newinliers) |
---|
102 | newoutliers = inliers(rejectIdx); |
---|
103 | |
---|
104 | %get new fund matrix |
---|
105 | %x1 = [f1(:, matches(1, newinliers))]; |
---|
106 | %x2 = [f2(:, matches(2, newinliers))]; |
---|
107 | %x = [x1; ones(1, length(x1)); x2; ones(1, length(x1))]; |
---|
108 | %F = fundmatrix(x); |
---|
109 | |
---|
110 | %re-evaluate all points |
---|
111 | %x1 = [f1(:, matches(1, :))]; |
---|
112 | %x2 = [f2(:, matches(2, :))]; |
---|
113 | %x = [x1; ones(1, length(x1)); x2; ones(1, length(x1))]; |
---|
114 | %refitinliers = getInliers(x, F, 0.00005); |
---|
115 | |
---|
116 | if (disp) |
---|
117 | % figure(3) |
---|
118 | % hist(sdist, 100) |
---|
119 | figure; |
---|
120 | vgg_gui_F(I1, I2, F'); |
---|
121 | |
---|
122 | figure; |
---|
123 | clf; |
---|
124 | plotmatches(I1,I2,f1, f2,matches(:, newinliers), 'Stacking', 'v', 'Interactive', 2) ; |
---|
125 | % figure(3) |
---|
126 | % clf; |
---|
127 | % plotmatches(I1,I2,f1, f2,matches(:, inliers), ... |
---|
128 | % 'Stacking', 'v', 'Interactive', 1, 'Dist', dist) ; |
---|
129 | %plotmatches(I1,I2,f1, f2,matches(:, refitinliers), 'Stacking', 'v', 'Interactive', 1) ; |
---|
130 | end |
---|
131 | end |
---|
132 | |
---|