基于双目系统的海面远距离测距研究OA北大核心CSTPCD

In order to adapt to the future development trend of unmanned ships and discover and locate long-range targets at sea,a binocular vision long-distance localization technology scheme is proposed by YOLOV8 and local stereo matching method.The binocular cameras are calibrated.The YOLOV8 algorithm is adopted to detect the targets at sea.Furthermore,by local stereo matching,the local feature extraction and local feature matching of the targets detected by the two cameras are carried out.On the basis of the matched results,the distance of the matched targets is calculated according to the principle of binocular vision.The experimental results show that the average accuracy of this technical solution for target ranging at distances of 100 to 300 m,300 to 500 m,500 to 700 m,and 700 to 900 m is 90%,84%,75%and 60%,respectively.Its processing speed is 20 PFS.In comparison with the traditional methods,the proposed method can focus on the targets at sea surface directly and implement stereo matching of local pixels only,so it has the advantages of fast operation speed and good positioning effect.