An autonomous sewer robots navigation based on stereo camera information

In this paper, we propose a method for autonomous sewer robots to navigate through a sewer pipe system based on stereo camera information. In this method, local features such as manholes and pipe joints are extracting as a feature pixels in the region of interest (ROI) of left image. Then, an accurate and fast stereo matching measure named linear computation is implemented in this ROI image to compute the distance between the robots and local features. Finally, the distance data can be used for navigation map in sewer pipe system. The experimental results show that our method can provide sufficient information for autonomous sewer robots navigation