Design And Implementation Of Target Detection System For USV-borne Marine Radar In Coastal Waters

Unmanned surface vehicle(USV)is an emerging maritime carrier.The accurate detection of obstacles on the sea is a prerequisite for USV to realize automatic obstacle avoidance.Compared with detection methods such as binocular vision and millimeter wave radar,marine radar has a wide detection range and can work around the clock.However,the existing methods are not accurate enough to detect sea surface obstacle targets,and it is impossible to guarantee the smooth progress of autonomous obstacle avoidance for USV.Aiming at the demand of USV in the coastal waters for accurate detection of sea surface obstacle targets,a set of USV-borne marine radar target detection systems is designed to achieve accurate detection of various sea surface obstacle targets in coastal waters.The specific research contents are as follows:Firstly,the functional requirements of the target detection system for marine radar is analyzed,the overall design of the detection system was completed,hardware selection and construction,designed and developed a communication interface module for the data communication between the detection system and the USV-borne TT&C computer.Obtain system detection results in real time on the TT&C computer,including parameters such as the location and movement of the sea surface obstacle targets,the latitude and longitude and heading of the USV,and radar images.By carrying out experimental tests,the effectiveness of the communication interface module is verified,and the accuracy of the system detection results is analyzed.Secondly,In order to enable the USV-borne TT&C computer to prioritize the detection results of the targets with a high degree of threat to the navigation safety of USV,a target detection method based on threat ranking is proposed,according to DCPA and TCPA between the target and the USV,comprehensively evaluate the threat level of each sea surface obstacle targets in the radar detection results to the navigation safety of the USV,and re-order them in descending order.In order to improve the real-time detection results of high-speed targets by the marine radar,a high-speed target detection method based on linear extrapolation position prediction is proposed.Predict the location of high-speed ship targets at multiple equal intervals in a radar scanning period,to supplement the radar detection results.In order to judge whether a large target is wrongly judged as multiple scattered targets by the marine radar,and describe the location of the target accurately,a large target detection method based on image contour features is proposed.Utilize the target contour in the radar image and the distance and orientation features contained in it,according to whether different targets correspond to the same contour in the image,analyze whether there is a misjudgment,use the closest distance of the target reflection section to the USV,and the azimuth of the two ends of the reflection section to the USV,to describe the location of the target.Finally,in order to verify the effectiveness and accuracy of the detection system,a application software with human-computer interaction,information storage,historical data query and other functions is designed and developed,and the software’s information flow model and functional structure.The actual ship experiment results in the coastal waters of Jiaozhou Bay show that the target detection system for USV-borne marine radar which is designed and developed in this thesis can effectively detect the sea surface obstacle targets.The proposed detection method makes the system detection results more accurate.