🟠 Point Cloud Points to Pixel Points
Function Description
This operator can use camera intrinsic matrix to project one or more three-dimensional spatial points onto the camera’s two-dimensional image plane, calculating the two-dimensional pixel coordinates corresponding to each three-dimensional point.
Usage Scenarios
-
3D feature point localization: After calculating coordinates of key points (such as object centers, pick points, corner points) in 3D space, this operator can be used to determine their exact pixel positions on 2D images.
-
2D result drawing and overlay: Visualize 3D calculation results on images. For example, project eight vertices of 3D detection boxes onto images, then draw 2D bounding boxes to verify the accuracy of 3D detection results.
-
By projecting vertices of ROI (regions of interest) in 3D space, dynamically generate corresponding pixel-level ROI on 2D images.
Input Output
Input |
Point cloud point list: A list containing N three-dimensional coordinate points with data shape N×3. Camera intrinsics: 3×3 camera intrinsic matrix. |
Output |
Pixel point list: A list containing N two-dimensional pixel coordinate points with data shape N×2, each element [u,v] corresponds to an input three-dimensional point. |