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
|
// ----------------------------------------------------------------------------
// - Open3D: www.open3d.org -
// ----------------------------------------------------------------------------
// Copyright (c) 2018-2024 www.open3d.org
// SPDX-License-Identifier: MIT
// ----------------------------------------------------------------------------
#include "open3d/pipelines/integration/UniformTSDFVolume.h"
#include <sstream>
#include "open3d/camera/PinholeCameraIntrinsic.h"
#include "open3d/data/Dataset.h"
#include "open3d/geometry/RGBDImage.h"
#include "open3d/io/ImageIO.h"
#include "open3d/utility/FileSystem.h"
#include "open3d/visualization/utility/DrawGeometry.h"
#include "tests/Tests.h"
namespace open3d {
namespace tests {
bool ReadPoses(const std::string& trajectory_path,
std::vector<Eigen::Matrix4d>& poses) {
FILE* f = utility::filesystem::FOpen(trajectory_path, "r");
if (f == NULL) {
utility::LogWarning("Read poses failed: unable to open file: {}",
trajectory_path);
return false;
}
char line_buffer[DEFAULT_IO_BUFFER_SIZE];
Eigen::Matrix4d pose;
auto read_pose = [&pose, &line_buffer, f]() -> bool {
// Read meta line
if (!fgets(line_buffer, DEFAULT_IO_BUFFER_SIZE, f)) {
return false;
}
// Read 4x4 matrix
for (size_t row = 0; row < 4; ++row) {
if (!fgets(line_buffer, DEFAULT_IO_BUFFER_SIZE, f)) {
return false;
}
if (sscanf(line_buffer, "%lf %lf %lf %lf", &pose(row, 0),
&pose(row, 1), &pose(row, 2), &pose(row, 3)) != 4) {
return false;
}
}
return true;
};
while (read_pose()) {
// Copy to poses
poses.push_back(pose);
}
fclose(f);
return true;
}
TEST(UniformTSDFVolume, Constructor) {
double length = 4.0;
int resolution = 128;
double sdf_trunc = 0.04;
auto color_type = pipelines::integration::TSDFVolumeColorType::RGB8;
pipelines::integration::UniformTSDFVolume tsdf_volume(
length, resolution, sdf_trunc,
pipelines::integration::TSDFVolumeColorType::RGB8);
// TSDFVolume base class attributes
EXPECT_EQ(tsdf_volume.voxel_length_, length / resolution);
EXPECT_EQ(tsdf_volume.sdf_trunc_, sdf_trunc);
EXPECT_EQ(tsdf_volume.color_type_, color_type);
// UniformTSDFVolume attributes
ExpectEQ(tsdf_volume.origin_, Eigen::Vector3d(0, 0, 0));
EXPECT_EQ(tsdf_volume.length_, length);
EXPECT_EQ(tsdf_volume.resolution_, resolution);
EXPECT_EQ(tsdf_volume.voxel_num_, resolution * resolution * resolution);
EXPECT_EQ(int(tsdf_volume.voxels_.size()), tsdf_volume.voxel_num_);
}
TEST(UniformTSDFVolume, RealData) {
// Poses
data::SampleRedwoodRGBDImages redwood_data;
std::string trajectory_path = redwood_data.GetOdometryLogPath();
std::vector<Eigen::Matrix4d> poses;
if (!ReadPoses(trajectory_path, poses)) {
throw std::runtime_error("Cannot read trajectory file");
}
// Extrinsics
std::vector<Eigen::Matrix4d> extrinsics;
for (const auto& pose : poses) {
extrinsics.push_back(pose.inverse());
}
// Intrinsics
camera::PinholeCameraIntrinsic intrinsic(
camera::PinholeCameraIntrinsicParameters::PrimeSenseDefault);
// TSDF init
pipelines::integration::UniformTSDFVolume tsdf_volume(
4.0, 100, 0.04, pipelines::integration::TSDFVolumeColorType::RGB8);
// Integrate RGBD frames
for (size_t i = 0; i < poses.size(); ++i) {
// Color
geometry::Image im_color;
io::ReadImage(redwood_data.GetColorPaths()[i], im_color);
// Depth
geometry::Image im_depth;
io::ReadImage(redwood_data.GetDepthPaths()[i], im_depth);
// Integrate
std::shared_ptr<geometry::RGBDImage> im_rgbd =
geometry::RGBDImage::CreateFromColorAndDepth(
im_color, im_depth, /*depth_scale*/ 1000.0,
/*depth_func*/ 4.0, /*convert_rgb_to_intensity*/ false);
tsdf_volume.Integrate(*im_rgbd, intrinsic, extrinsics[i]);
}
// These hard-coded values are for unit test only. They are used to make
// sure that after code refactoring, the numerical values still stay the
// same. However, using different parameters or algorithmtic improvements
// could invalidate these reference values. We use a custom threshold 0.1
// to account for acccumulative floating point errors.
// Extract mesh
std::shared_ptr<geometry::TriangleMesh> mesh =
tsdf_volume.ExtractTriangleMesh();
EXPECT_EQ(mesh->vertices_.size(), 3198u);
EXPECT_EQ(mesh->triangles_.size(), 4402u);
Eigen::Vector3d color_sum(0, 0, 0);
for (const Eigen::Vector3d& color : mesh->vertex_colors_) {
color_sum += color;
}
ExpectEQ(color_sum, Eigen::Vector3d(2703.841944, 2561.480949, 2481.503805),
/*threshold*/ 0.1);
// Uncomment to visualize
// visualization::DrawGeometries({mesh});
// Extract point cloud
std::shared_ptr<geometry::PointCloud> pcd = tsdf_volume.ExtractPointCloud();
EXPECT_EQ(pcd->points_.size(), 2227u);
EXPECT_EQ(pcd->colors_.size(), 2227u);
color_sum << 0, 0, 0;
for (const Eigen::Vector3d& color : pcd->colors_) {
color_sum += color;
}
ExpectEQ(color_sum, Eigen::Vector3d(1877.673116, 1862.126057, 1862.190616),
/*threshold*/ 0.1);
Eigen::Vector3d normal_sum(0, 0, 0);
for (const Eigen::Vector3d& normal : pcd->normals_) {
normal_sum += normal;
}
ExpectEQ(normal_sum, Eigen::Vector3d(-161.569098, -95.969433, -1783.167177),
/*threshold*/ 0.1);
// Extract voxel cloud
std::shared_ptr<geometry::PointCloud> voxel_pcd =
tsdf_volume.ExtractVoxelPointCloud();
EXPECT_EQ(voxel_pcd->points_.size(), 4488u);
EXPECT_EQ(voxel_pcd->colors_.size(), 4488u);
color_sum << 0, 0, 0;
for (const Eigen::Vector3d& color : voxel_pcd->colors_) {
color_sum += color;
}
ExpectEQ(color_sum, Eigen::Vector3d(2096.428416, 2096.428416, 2096.428416),
/*threshold*/ 0.1);
}
} // namespace tests
} // namespace open3d
|
