Validation of Pose Correction Based on Photometric Error Minimization Using LiDAR Reflectance Intensity Images Involving Missing Data in Degenerate Environments

Abstract

This report presents a validation of a pose correction method based on minimizing photometric error in LiDAR reflectance intensity images, even when these images contain missing data due to low-density point clouds in degenerate environments. Previous approaches have used reflectance intensity images generated from LiDAR point clouds to correct the pose; however, these methods typically require high-density point clouds, and significant missing data can occur when the density is low. In this work, we address that limitation by employing a photometric error minimization approach, enabling robust pose correction even in the presence of incomplete reflectance intensity images. The effectiveness of the proposed method was verified using data acquired in a laboratory environment.