🟠 Point Cloud Bounding Box to 2D Bounding Box

Function Description

This operator is used to convert three-dimensional point cloud bounding boxes to two-dimensional pixel bounding boxes in images through camera intrinsic matrix projection transformation, and simultaneously generate corresponding two-dimensional ROI regions.

Usage Scenarios

Commonly used in vision and point cloud result fusion, detection result visualization scenarios, to annotate regions corresponding to point cloud bounding boxes on images.

Input Output

Input

Point cloud bounding box: Three-dimensional bounding box in camera coordinate system.

Camera intrinsics: 3×3 camera intrinsic matrix.

Image (optional): Optional RGB image for visualization output.

Output

Pixel bounding box: Four-point pixel boxes corresponding to each point cloud bounding box, format N×4×2.

2D ROI: Rectangular regions corresponding to each bounding box.