Performance characterization of a high-speed stereo vision sensor for acquisition of time-varying 3D shapes

摘要

Acquisition of dynamic dense 3D shape data is of increasing importance in computer vision with applications in various disciplines. In this paper, we investigate the performance of a unique high-speed range sensor based on the stereo vision principle for 3D shape acquisition of animals. The investigation reveals some characteristics of the current version of the sensor with respect to its physical parameters, which suggest an more appropriate configuration of the sensor in real data acquisition scenarios. Due to the novelty of the sensor and the application, we believe that our evaluation of the sensor's performance will inspire new applications to follow using the dynamic 3D acquisition technology of similar types.