PRISM: Color-Stratified Point Cloud Sampling

TL;DR

PRISM is a novel color-guided stratified sampling method for RGB-LiDAR point clouds that preserves textured regions by allocating samples based on chromatic diversity, improving 3D reconstruction.

Contribution

It introduces a new sampling approach that considers color diversity, shifting from spatial uniformity to visual complexity for better feature preservation.

Findings

Preserves texture-rich regions with high color variation.

Reduces samples in homogeneous, repetitive areas.

Enhances 3D reconstruction quality with sparser point clouds.

Abstract

We present PRISM, a novel color-guided stratified sampling method for RGB-LiDAR point clouds. Our approach is motivated by the observation that unique scene features often exhibit chromatic diversity while repetitive, redundant features are homogeneous in color. Conventional downsampling methods (Random Sampling, Voxel Grid, Normal Space Sampling) enforce spatial uniformity while ignoring this photometric content. In contrast, PRISM allocates sampling density proportional to chromatic diversity. By treating RGB color space as the stratification domain and imposing a maximum capacity k per color bin, the method preserves texture-rich regions with high color variation while substantially reducing visually homogeneous surfaces. This shifts the sampling space from spatial coverage to visual complexity to produce sparser point clouds that retain essential features for 3D reconstruction tasks.