Dynamic Sensor Matching for Parallel Point Cloud Data Acquisition

Based on depth perception of individual stereo cameras, spatial structures can be derived as point clouds. The quality such three-dimensional data is technically restricted by sensor limitations, latency recording, and insufficient object reconstructions caused surface illustration. Additionally external physical effects like lighting conditions, material properties, reflections lead to deviations between real virtual perception. Such influences seen in rendered clouds geometrical imaging errors surfaces edges. We propose the simultaneous use multiple dynamically arranged cameras. increased information density leads more details surrounding detection During a pre-processing phase collected are merged prepared. Subsequently, logical analysis part examines allocates captured images space. For this purpose, it necessary create new metadata set consisting image localisation data. post-processing reworks matches locally assigned images. As result, dynamic moving become comparable so that accurate cloud generated. evaluation better comparability we decided synthetically generated sets. Our approach builds foundation for real-time based generation digital twins with aid