Overview of RGBD semantic segmentation based on deep learning

Semantic segmentation is one of the basic tasks in computer vision. Its purpose is to achieve pixel-level scene segmentation. With the popularity of depth sensors, combining depth data with RGB images for semantic segmentation can improve the accuracy of semantic segmentation. First, this paper mainly summarizes the fusion of RGB information and depth information and then describes the RGBD semantic segmentation method, evaluation metrics, data set, and comparison of the results on the two mainstream data sets, and then make a prospect of possible future research directions, and finally, a conclusion is made. This part of the work has a certain guiding significance for future research on RGBD semantic segmentation and lays a foundation for later research.