Real-time pattern recognition for guidance of an autonomous undersea submersible

Abstract

This paper reports the initial results of an effort to develop simple and fast vision algorithms on compact and imoeddaDle hardware for the guidance and control of an autonomous underwater vehicle. The specific application involves tracking underwater cables and chains. Feature points are identified in the underwater video images using a technique which combines segmentation by gray level and run length. Hough transformation is then used to find the straight line in the image. The process is performed at a throughput of approximately 1 image per second using a PC-bus video frame grabber and a PC/AT compatible micro-computer.