A real-time system for 3D recovery of dynamic scenes based on multiple RGBD imagers

This paper presents a real-time system for 3D recovery of dynamic scenes by using multiple FPGA-based RGBD imagers. The RGBD imager developed in our lab produces color images combined with the corresponding dense disparity maps encoding depth information. Multiple RGBD imagers were externally triggered to sense the 3D world synchronously. To improve the quality of the acquired depth map, a probabilistic method for fusing depth maps from multiple RGBD imagers with sparsely distributed viewpoints was proposed. The fused depth maps of a dynamic scene from multiple viewpoints were then streamed to a PC cluster to generate the 3D point cloud of the dynamic scene in real time. Experimental results show that our system is very promising for the 3D recovery of large scale dynamic scenes with multiple moving objects.