Autonomous obstacle avoidance assistant system for unmanned surface vehicle based on Intelligent Vision
Zhihong Xu, Kunpeng Duan, Dongzhe Li · 2022 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · 2022
In view of the current situation of ship borne automatic identification system, automatic radar and vision based technology applied in the field of ship collision avoidance at home and abroad, there are common shortcomings, that is, it is difficult to accurately identify in the poor navigation environment with low visibility such as night and fog, or in the complex navigation environment with dense targets such as narrow channel and heavy traffic, and the contrast of collected image information is lowDue to the problems of low signal-to-noise ratio and easy missing detection and loss, the intelligent obstacle avoidance system of unmanned surface vehicle based on intelligent vision constructed in this project can not only overcome the adverse environmental factors of fog and improve the perception accuracy;It can also solve the detection problem when the targets are dense and the navigation environment is complex, and obtain more intuitive external navigation environment information.The intelligent obstacle avoidance system of unmanned surface vehicle based on intelligent vision constructed in this project is mainly composed of three parts: visual perception enhancement, obstacle target autonomous recognition and unmanned boat autonomous obstacle avoidance.Firstly, the obtained video image is processed;Secondly, yolov4 target detection network is used to identify and enhance obstacles in video images.Then, the obstacle blocking angle in the image is obtained by computer technology, and the obstacle distance is measured by millimeter wave radar to realize the fusion of vision and radar information.Finally, the intelligent collision avoidance of unmanned craft is realized based on VFH obstacle avoidance algorithm.While solving the problems of the current ship borne identification system, it enhances the type identification effect of obstacles, realizes the effect of accurate identification, and greatly reduces the impact of external interference factors on the safe navigation of unmanned craft. It has high application value and prospect.