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
|
function [v1, v2, cam_offsets, t, R ] = createMulti2D2DOmniExperiment( pt_per_cam, cam_number, noise, outlier_fraction )
%% generate the camera system
cam_distance = 1.0;
%% set a regular camera system with 2 or 4 cameras here
if cam_number == 2
cam_offsets = [ cam_distance -cam_distance; 0.0 0.0; 0.0 0.0 ];
else
cam_number = 4; % only two or 4 supported for this experiment
cam_offsets = [ cam_distance 0.0 -cam_distance 0.0; 0.0 cam_distance 0.0 -cam_distance; 0.0 0.0 0.0 0.0 ];
end
%% generate random view-points
max_parallax = 2.0;
max_rotation = 0.5;
position1 = zeros(3,1);
rotation1 = eye(3);
position2 = max_parallax * 2.0 * (rand(3,1) - repmat(0.5,3,1));
rotation2 = generateBoundedR(max_rotation);
%% Generate random point-cloud
avg_depth_over_cam_distance = 10.0;
maxDepth = 5.0;
p = cell([cam_number 1]);
for cam=1:cam_number
normalizedPoints = 2.0*(rand(3,pt_per_cam)-repmat(0.5,3,pt_per_cam));
p{cam,1} = maxDepth * normalizedPoints;
end
%% Now create the correspondences by looping through the cameras
focal_length = 800.0;
v1 = cell([cam_number 1]);
v2 = cell([cam_number 1]);
for cam=1:cam_number
v1{cam,1} = zeros(3,pt_per_cam);
v2{cam,1} = zeros(3,pt_per_cam);
for i=1:pt_per_cam
cam_offset = cam_offsets(:,cam);
%special: shift the point in the first frame along current camera axis, which guarantees homogeneous distribution
temp = p{cam,1}(:,i) + avg_depth_over_cam_distance * cam_offset;
p{cam,1}(:,i) = temp;
body_point1 = rotation1' * (p{cam,1}(:,i)-position1);
body_point2 = rotation2' * (p{cam,1}(:,i)-position2);
% we actually omit the can rotation here by unrotating the bearing
% vectors already
bearingVector1 = body_point1 - cam_offset;
bearingVector2 = body_point2 - cam_offset;
bearingVector1_norm = norm(bearingVector1);
bearingVector2_norm = norm(bearingVector2);
bearingVector1 = bearingVector1/bearingVector1_norm;
bearingVector2 = bearingVector2/bearingVector2_norm;
% add noise to the bearing vectors here
bearingVector1_noisy = addNoise(bearingVector1,focal_length,noise);
bearingVector2_noisy = addNoise(bearingVector2,focal_length,noise);
% store the normalized bearing vectors along with the cameras they are
% being seen (we create correspondences that always originate from the
% same camera, you can change this if you want)
bearingVector1_norm = norm(bearingVector1_noisy);
bearingVector2_norm = norm(bearingVector2_noisy);
v1{cam,1}(:,i) = [bearingVector1_noisy./bearingVector1_norm];
v2{cam,1}(:,i) = [bearingVector2_noisy./bearingVector2_norm];
end
end
%% Add outliers
outliers_per_cam = floor(outlier_fraction*pt_per_cam);
if outliers_per_cam > 0
for cam=1:cam_number
for i=1:outliers_per_cam
cam_offset = cam_offsets(:,cam);
%generate random point
normalizedPoint = 2.0*(rand(3,1)-repmat(0.5,3,1));
point = maxDepth * normalizedPoint + avg_depth_over_cam_distance * cam_offset;
body_point2 = rotation2' * (point-position2);
% store the point (no need to add noise)
bearingVector2 = body_point2 - cam_offset;
% store the normalized bearing vectors along with the cameras they are
% being seen (we create correspondences that always originate from the
% same camera, you can change this if you want)
bearingVector2_norm = norm(bearingVector2);
v2{cam,1}(:,i) = [bearingVector2./bearingVector2_norm];
end
end
end
%% compute relative translation and rotation
R = rotation1' * rotation2;
t = rotation1' * (position2 - position1);
|
