How phantom0122/3d_object_reconstruction works
This repository implements multi-view RGB-D object reconstruction: capture and align color/depth frames, extract object point clouds, estimate pairwise poses, optimize a pose graph, and produce a combined reconstruction. Its documented registration options include SIFT, ORB, LoFTR, and point-to-plane ICP. The graph also shows the feature-matching and geometry utility layer. Some stages are documented more fully than their sampled implementation, so those links cite the README; no unsampled implementation details are assumed.
Main entry points: pose_graph_Feature_based.py Feature pose graph · pcd_o3d Object point clouds · SIFT.py SIFT matching · registration.py Registration methods
Components 17
RGB-D capture
- Aligned capture
capture_aligned_images.py - Aligned RGB-D frames
train
Point cloud preparation
- Object extraction
preprocess_pcd.py - RGB-D conversion
RGBD_to_pcd.py - Object point clouds
pcd_o3d
Pairwise registration
- SIFT matching
SIFT.py - ORB matching
ORB.py - LoFTR matching
LoFTR.py - Registration methods
registration.py
Reconstruction
- Feature pose graph
pose_graph_Feature_based.py - ICP pose graph
pose_graph_ICP.py - Reconstructed clouds
results
Internal support
- Registration visualization
plot.py - Geometry utilities
utils.py - Cloud depth rendering
pcd_to_depth.py
Other components
- Operator
- RealSense camera
Connections 23
- Operator → Aligned capture: starts capture
README.md - RealSense camera → Aligned capture: streams frames
capture_aligned_images.py - Aligned capture → Aligned RGB-D frames: saves aligned frames
README.md - Aligned RGB-D frames → Object extraction: provides RGB-D
preprocess_pcd.py - Object extraction → Object point clouds: writes clouds
preprocess_pcd.py - Aligned RGB-D frames → RGB-D conversion: provides RGB-D
- RGB-D conversion → Object point clouds: writes cloud
- Object point clouds → Feature pose graph: registers views
README.md - Object point clouds → ICP pose graph: registers views
README.md - Feature pose graph → SIFT matching: calls loop closure
pose_graph_Feature_based.py - Feature pose graph → ORB matching: selects matcher
pose_graph_Feature_based.py - Feature pose graph → LoFTR matching: selects matcher
pose_graph_Feature_based.py - ICP pose graph → SIFT matching: calls loop closure
pose_graph_ICP.py - SIFT matching → Registration methods: uses RANSAC
SIFT.py - ORB matching → Registration methods: uses RANSAC
ORB.py - LoFTR matching → Registration methods: uses RANSAC
LoFTR.py - SIFT matching → Geometry utilities: gets boundaries
SIFT.py - ORB matching → Geometry utilities: gets boundaries
ORB.py - LoFTR matching → Geometry utilities: gets boundaries
LoFTR.py - Registration methods → Registration visualization: draws alignment
registration.py - Feature pose graph → Reconstructed clouds: produces reconstruction
README.md - ICP pose graph → Reconstructed clouds: produces reconstruction
README.md - Object point clouds → Cloud depth rendering: renders depth