High quality depth maps are required in various applications including robotics, computer vision, and autonomous driving. In order to obtain precise depth information, there has been an increasing interest in the combination of different depth sensors...
High quality depth maps are required in various applications including robotics, computer vision, and autonomous driving. In order to obtain precise depth information, there has been an increasing interest in the combination of different depth sensors. Depth information can be acquired real-time by Time-of-Flight (ToF) cameras and stereo cameras. There are various Depth-Stereo fusing methods to overcome the limitations of a single depth sensor, including Conditional Cost Volume Normalization algorithm. The CCVNorm framework is generic and closely integrated with the cost volume component that is commonly utilized in stereo matching neural networks. In this paper, we give an overview over methods for the fusion of depth and passive stereo data. We experimentally evaluate the performance and robustness of various Depth-Stereo fusion methods with the KITTI Stereo and Depth Completion datasets.