A study on plane extraction from distance images using 3D Hough transform

Studies have been conducted to detect object planes in 3D using 3D distant measurement camera such as Swiss Ranger SR-4000. In those studies, first, objects in the gray scale range image from the camera are emphasized. Then, emphasized image is grouped regarding the depth information and parameters for planes are calculated using least-square-method. Here, emphasis is sensitive enough, but ambiguity of gray scale range image causes error detections. Considering this problem, a new plane detection method is proposed using 3D Hough transform. According to the experiments, new proposal is effective and it showed better results compared to the previous methods.