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
|
// ----------------------------------------------------------------------------
// - Open3D: www.open3d.org -
// ----------------------------------------------------------------------------
// Copyright (c) 2018-2024 www.open3d.org
// SPDX-License-Identifier: MIT
// ----------------------------------------------------------------------------
#include "open3d/t/pipelines/slam/Frame.h"
#include "open3d/t/pipelines/slam/Model.h"
#include "pybind/docstring.h"
namespace open3d {
namespace t {
namespace pipelines {
namespace slam {
static const std::unordered_map<std::string, std::string>
map_shared_argument_docstrings = {
{"voxel_size", "The voxel size of the volume in meters."},
{"block_resolution",
"Resolution of local dense voxel blocks. By default 16 "
"is used to create 16^3 voxel blocks."},
{"block_count",
"Number of estimate blocks per scene with the block "
"resolution set to 16 and the 6mm voxel resolution. "
"Typically 20000 for small scenes (a desk), 40000 for medium "
"scenes (a bedroom), 80000 for large scenes (an "
"apartment)."},
{"transformation", "A 4x4 3D transformation matrix."},
{"device", "The CPU or CUDA device used for the object."},
{"depth_max",
"The max clipping depth to filter noisy observations too "
"far."},
{"depth_min", "The min clipping depth."},
{"depth_scale",
"The scale factor to convert raw depth into meters."},
{"input_frame",
"The frame that contains raw depth and optionally images with "
"the same size from the input."},
{"model_frame",
"The frame that contains ray casted depth and optionally "
"color from the volumetric model."},
{"estimated_number",
"Estimated number of surface points. Use -1 if no estimation "
"is available."},
{"weight_threshold",
"Weight threshold to filter outlier voxel blocks."},
{"height", "Height of an image frame."},
{"width", "Width of an image frame."},
{"intrinsics", "Intrinsic matrix stored in a 3x3 Tensor."},
{"trunc_voxel_multiplier",
"Truncation distance multiplier in voxel size for signed "
"distance. For instance, "
"--trunc_voxel_multiplier=8 with --voxel_size=0.006(m) "
"creates a truncation distance of 0.048(m)."}};
void pybind_slam_declarations(py::module &m) {
py::module m_slam = m.def_submodule("slam", "Tensor DenseSLAM pipeline.");
py::class_<Model> model(m_slam, "Model",
"Volumetric model for Dense SLAM.");
py::class_<Frame> frame(m_slam, "Frame",
"A frame container that stores a map from keys "
"(color, depth) to tensor images.");
}
void pybind_slam_definitions(py::module &m) {
auto m_slam = static_cast<py::module>(m.attr("slam"));
auto model = static_cast<py::class_<Model>>(m_slam.attr("Model"));
py::detail::bind_copy_functions<Model>(model);
model.def(py::init<>());
model.def(py::init<float, int, int, core::Tensor, core::Device>(),
"Constructor of a VoxelBlockGrid", "voxel_size"_a,
"block_resolution"_a = 16, " block_count"_a = 10000,
"transformation"_a = core::Tensor::Eye(4, core::Float64,
core::Device("CPU:0")),
"device"_a = core::Device("CUDA:0"));
docstring::ClassMethodDocInject(m_slam, "Model", "__init__",
map_shared_argument_docstrings);
model.def("get_current_frame_pose", &Model::GetCurrentFramePose);
model.def("update_frame_pose", &Model::UpdateFramePose);
model.def("synthesize_model_frame", &Model::SynthesizeModelFrame,
py::call_guard<py::gil_scoped_release>(),
"Synthesize frame from the volumetric model using ray casting.",
"model_frame"_a, "depth_scale"_a = 1000.0, "depth_min"_a = 0.1,
"depth_max"_a = 3.0, "trunc_voxel_multiplier"_a = 8.0,
"enable_color"_a = false, "weight_threshold"_a = -1.0);
docstring::ClassMethodDocInject(m_slam, "Model", "synthesize_model_frame",
map_shared_argument_docstrings);
model.def(
"track_frame_to_model", &Model::TrackFrameToModel,
py::call_guard<py::gil_scoped_release>(),
"Track input frame against raycasted frame from model.",
"input_frame"_a, "model_frame"_a, "depth_scale"_a = 1000.0,
"depth_max"_a = 3.0, "depth_diff"_a = 0.07,
py::arg_v("method", odometry::Method::PointToPlane,
"Method.PointToPlane"),
"criteria"_a = (std::vector<odometry::OdometryConvergenceCriteria>){
6, 3, 1});
docstring::ClassMethodDocInject(m_slam, "Model", "track_frame_to_model",
map_shared_argument_docstrings);
model.def("integrate", &Model::Integrate,
py::call_guard<py::gil_scoped_release>(),
"Integrate an input frame to a volume.", "input_frame"_a,
"depth_scale"_a = 1000.0, "depth_max"_a = 3.0,
"trunc_voxel_multiplier"_a = 8.0);
docstring::ClassMethodDocInject(m_slam, "Model", "integrate",
map_shared_argument_docstrings);
model.def("extract_pointcloud", &Model::ExtractPointCloud,
py::call_guard<py::gil_scoped_release>(),
"Extract point cloud from the volumetric model.",
"weight_threshold"_a = 3.0, "estimated_number"_a = -1);
docstring::ClassMethodDocInject(m_slam, "Model", "extract_pointcloud",
map_shared_argument_docstrings);
model.def("extract_trianglemesh", &Model::ExtractTriangleMesh,
py::call_guard<py::gil_scoped_release>(),
"Extract triangle mesh from the volumetric model.",
"weight_threshold"_a = 3.0, "estimated_number"_a = -1);
docstring::ClassMethodDocInject(m_slam, "Model", "extract_trianglemesh",
map_shared_argument_docstrings);
model.def(
"get_hashmap", &Model::GetHashMap,
"Get the underlying hash map from 3D coordinates to voxel blocks.");
model.def_readwrite("voxel_grid", &Model::voxel_grid_,
"Get the maintained VoxelBlockGrid.");
model.def_readwrite("frustum_block_coords", &Model::frustum_block_coords_,
"Active block coordinates from prior integration");
model.def_readwrite("transformation_frame_to_world",
&Model::T_frame_to_world_,
"Get the 4x4 transformation matrix from the current "
"frame to the world frame.");
model.def_readwrite("frame_id", &Model::frame_id_,
"Get the current frame index in a sequence.");
auto frame = static_cast<py::class_<Frame>>(m_slam.attr("Frame"));
py::detail::bind_copy_functions<Frame>(frame);
frame.def(py::init<int, int, core::Tensor, core::Device>(), "height"_a,
"width"_a, "intrinsics"_a, "device"_a);
docstring::ClassMethodDocInject(m_slam, "Frame", "__init__",
map_shared_argument_docstrings);
frame.def("height", &Frame::GetHeight);
frame.def("width", &Frame::GetWidth);
frame.def("set_data", &Frame::SetData,
"Set a 2D tensor to a image to the given key in the map.");
frame.def("get_data", &Frame::GetData,
"Get a 2D tensor from a image from the given key in the map.");
frame.def("set_data_from_image", &Frame::SetDataFromImage,
"Set a 2D image to the given key in the map.");
frame.def("get_data_as_image", &Frame::GetDataAsImage,
"Get a 2D image from from the given key in the map.");
}
} // namespace slam
} // namespace pipelines
} // namespace t
} // namespace open3d
|
