Abstract

Continuous-wave Time-of-flight (TOF) range imaginghas become a commercially viable technology with manyapplications in computer vision and graphics. However, thedepth images obtained from TOF cameras contain scene de-pendent errors due to multipath interference (MPI). Specifi-cally, MPI occurs when multiple optical reflections return toa single spatial location on the imaging sensor. Many priorapproaches to rectifying MPI rely on sparsity in optical re-flections, which is an extreme simplification. In this paper, we correct MPI by combining the standard measurementsfrom a TOF camera with information from direct and global light transport. We report results on both simulated experiments and physical experiments (using the Kinect sensor). Our results, evaluated against ground truth, demonstrate a quantitative improvement in depth accuracy.