Compensation of Beam Divergence Effects in Multimodal Sensor Fusion

Abstract

Greg Passmore

Misalignment among optical, infrared, LiDAR, and radar systems arises from fundamental differences in their sampling geometries and divergence characteristics rather than from calibration error. Each modality observes the environment through a distinct sampling function: passive systems integrate radiance over an angular cone, while active systems record discrete returns from narrow beams. As range increases, these differences produce nonlinear distortions, occlusion inconsistencies, and radiometric bias that cannot be resolved through geometric registration alone. This paper formalizes the physical basis of divergence- induced misalignment and introduces a unified correction framework implemented through the Vossel-Aligned Spatial Lattice (VASL). VASL embeds all sensor data within a volumetric lattice where divergence, incidence angle, and occlusion are locally compensated through range-aware kernels and uncertainty propagation. The method includes equirectangular reprojection, anisotropic footprint modeling, occlusion-aware radiometric fusion, and visibility weighting. This unified framework corrects divergence and material interaction effects across active and passive domains, enabling coherent multimodal alignment for analytical fusion and machine learning applications.

PDF