Dense Depth Map Acquisition System for 3DTV Applications Based on Active Stereo and Structured Light Integration

Abstract

In this paper, we present and analyze a depth imaging system based on the integration of active stereo matching and structured light methods. The integration of these methods benefits from the advantages of the two approaches, allowing a shape recovery from a wider view with less occlusion. We build a system composed of two cameras and a projector, and project a single one-shot pattern. We first use the structured light part in order to estimate reliable correspondences between each camera and the projector via an efficient pattern decoding technique. The remaining unresolved regions are explored by a stereo matching technique which is less sensitive to object surface colors and projectors short depth of field to estimate additional correspondences. By switching between the colored pattern and a white light, the texture information of the system is retrieved at the same time and from the same viewpoint. Finally, we present a thoughtful and in-depth analysis of the capabilities and limitations of the presented system in the context of the development and contents creation for depth image-based representation (DIBR) 3DTV. Through carefully designed experiments we quantify the depth range of the camera system, the effects of the projector depth of field on the pattern decoding performance and the robustness to scene surface colors with respect to their hue, saturation, and brightness.