LiDAR-Camera Fusion¶

lidar_camera_fusion: Sensor fusion operations: distance-based colorization, region-of-interest extraction, and multi-resolution analysis for combined LiDAR and camera processing.


Methods	`intensity_colorization`, `roi_extraction`, `multi_resolution`
Input files	`.ply`, `.pcd`, `.txt`, `.xyz`, `.las`
Bundled datasets to try	`open3d_bunny`, `open3d_icp_demo`, `kitti_raw_2011_09_26_drive_0001`, `medshapenet`, `c3vd_endoscopy_splats`
CLI	`pardesline run lidar_camera_fusion.<method> input.ext [--param value]`

How it works¶

LiDAR-Camera fusion combines 3D point cloud data with 2D imagery. A 3D point \(P = (X,Y,Z)\) projects to pixel \((u,v)\) via:

\[ \begin{pmatrix} u \\ v \\ 1 \end{pmatrix} = \frac{1}{Z} K [R|t] \begin{pmatrix} X \\ Y \\ Z \\ 1 \end{pmatrix} \]

Methods¶

intensity_colorization¶

Points colored by distance from origin using a selected colormap.

import pardesline as pl

obj = pl.load("input.ply")
result = obj.run("lidar_camera_fusion.intensity_colorization")  # catalog defaults
result.provenance[-1]  # the step, every parameter used, duration, version

Output: .ply. Script: python -m pardesline.ops.lidar_camera_fusion.intensity_colorization --input ... --output ....

Defaults validated on: KITTI sweeps carry reflectance in the fourth column (manifest layout); generated clouds in tests/test_ops_smoke.py

Parameter	Type	Default	Range	What it does	Why this default, when to change it
`colormap`	int	`0`	0 to 3	0=jet, 1=viridis, 2=inferno, 3=coolwarm	0 (jet) for reflectance: lane paint and plates saturate to red, asphalt stays blue, which is what people look for in an intensity view.

roi_extraction¶

Extract a sub-volume defined by axis-aligned percentage bounds.

import pardesline as pl

obj = pl.load("input.ply")
result = obj.run("lidar_camera_fusion.roi_extraction")  # catalog defaults
result.provenance[-1]  # the step, every parameter used, duration, version

Output: .ply. Script: python -m pardesline.ops.lidar_camera_fusion.roi_extraction --input ... --output ....

Defaults validated on: generated clouds in tests/test_ops_smoke.py

Parameter	Type	Default	Range	What it does	Why this default, when to change it
`x_range`	float	`0.5`	0.0 to 1.0	% of X-range to keep (centered).	Keep the central 50 % along x: the near field in front of and behind the sensor, where the camera overlap is.
`y_range`	float	`0.5`	0.0 to 1.0	% of Y-range to keep (centered).	Keep the central 50 % along y: the lane and its neighbours.
`z_range`	float	`0.8`	0.0 to 1.0	% of Z-range to keep (centered).	Keep 80 % along z: drops only the extreme returns (tree tops, below-ground reflections).

multi_resolution¶

Near regions use fine voxels, far regions use coarse voxels.

import pardesline as pl

obj = pl.load("input.ply")
result = obj.run("lidar_camera_fusion.multi_resolution")  # catalog defaults
result.provenance[-1]  # the step, every parameter used, duration, version

Output: .ply. Script: python -m pardesline.ops.lidar_camera_fusion.multi_resolution --input ... --output ....