Research On Underwater Optical Target Detection And Tracking Technology Based On Binocular Vision

As a big ocean country,development and utilization of marine resources has important military and economic value for the China.Underwater vehicles are an important tool for the detection,development and utilization of marine resources.Bionic underwater vehicles have attracted great attention in industry and academia in recent years due to their great advantages in maneuverability,energy consumption and concealment.In near-water and shallow water environments,stereo vision sensors can be used to obtain more abundant image information than traditional acoustic equipment,and then can improve the perception and autonomous navigation capability of the aircraft in complex location environments.However,there are some difficult problems such as poor image quality,environmental sensitivity,and low robustness in underwater vision technology applying.In this regard,with the support of the Science and Technology Commission of the Military Commission and the Shipbuilding Engineering Research Institute project,this paper focuses on the research of optical target detection and tracking and experimental platform construction of underwater bionic aircraft.The specific research contents are as follows.Aiming at underwater vision technology,this paper takes laser beam as the research target,analyzes its underwater imaging characteristics,enhances target contrast by color feature and geometric feature fusion,and target detection done by morphological operation and target straight line fitting,In the curve fitting part,the robustness of detection can be improved by the method of up and down search.Under the target feature point changes greatly and the single underwater background,the Kalman filtering technique based on information fusion is studied to solve the target tracking problem.At the same time,the motion template tracking algorithm is optimized for specific targets to improve the tracking effect.Target tracking performance was verified and evaluated through laboratory test.The difference between camera models in underwater and air is analyzed,and the relationship between underwater target and image distortion caused by different positions of the camera is derived.The calibration of indoor and underwater camera calibration is carried out.Based on this,a binocular vision model is built to solve the spatial position of the target relative to the camera.Through the laboratory test,the accuracy of the solution of the spatial position is verified and evaluated.The bionic underwater vehicle that relies on the BCF movement mode(such as the black fish type)was designed and manufactured.The underwater tests show that the bionic fish has good maneuverability,and the actual turning radius is shorter than body length;the swimming speed meets the project requirements,but there is plenty of room for improvement.The outdoor underwater test and target tracking test of the blackfish-like aircraft were carried out.On the aircraft platform,the code realization and module test of the vision-based sensing and control algorithm are completed,and the swimming test of the underwater laser beam target is monitored by underwater binocular vision detection.Based on the measured data,the accuracy and robustness of the control algorithm,target position solving model and tracking algorithm are evaluated.