A method for object detection using point cloud measurement in the sea environment

This paper describes a method for detection of object using 3D point cloud measurement in the sea environment. The method employs RBNN clustering method and using a 3D Lidar, mono-vision and stereo-vision cameras, and radar vision system. A radially based nearest neighbors (RBNN) clustering technique is adopted to perform object detection on 3D point cloud clustering. RBNN is constructing clusters based on the radius or distance parameter. In RBNN, each 3D point searches its nearest neighbor (NN) under some radius threshold value and combines all the neighboring points as a group or cluster. The experimental results verify the performance of RBNN to detect objects from 3D point cloud measurements in sea environment.