A Research On Panoramic Surround View System Of Unmanned Surface Vessel

Panoramic surround view system is often used in driving assistance systems to reduce traffic accident rate,precision parking and other safe driving problems in the automotive industry,but in recent years,with the development of economy and society,the exploration of marine resources has become an inevitable trend,so it has gradually begun to be used in the ship industry,unmanned ship direction is one of the hot directions.In the case of controlling unmanned ships to carry out various operations,the controller standing on the shore does not know the specific situation of the water,but the 2D bird’s eye view generated by the panoramic surround view system can provide large-field monitoring,and its wide viewing angle provides convenience for monitoring the surrounding environment.Therefore,it is necessary to study the application of panoramic surround view system in unmanned ship platforms.Most of the panoramic surround view systems in the automotive industry are implemented based on image stitching algorithms based on feature point matching.However,in the field of unmanned ships,due to the particularity of the water surface environment,it is difficult to find a sufficient number of overlapping parts of the images captured by the fisheye lens.Therefore,this paper makes some improvements to the panoramic view system widely used in the field of smart cars to adapt to the environment of unmanned ship operations.This paper conducts corresponding research on the main process of the panoramic view system of unmanned ships,and the main work is as follows:(1)Distortion correction of fisheye lens.The imaging model of the fisheye lens is analyzed,the principle of Zhang Zhengyou’s calibration algorithm is studied,and Zhang Zhengyou’s calibration algorithm is tested using a calibration board.(2)The stitching problem of panoramic images.First of all,two types of image stitching algorithms are studied,the first type is image stitching algorithm based on feature point matching,and the other type is image stitching algorithm that does not need feature point matching.According to the characteristics of unmanned ships,a unified world coordinate system method that does not require feature point matching is used for image stitching operations instead of an image stitching algorithm based on feature point matching.Secondly,three image fusion algorithms are studied,and finally the weighted average method is used to eliminate Problems with "seamlines" created during the splicing process.(3)The enhancement of panoramic images.Although the image fused by the weighted average method eliminates the problem of "seamline",the brightness and contrast of the fused result image are low.The final image enhancement can effectively improve the brightness and contrast of the image and improve the visual effect.Finally,the overall scheme of the panoramic view system for unmanned ships is given.According to the image correlation algorithm studied earlier,the overall algorithm flow is designed,and the corresponding simulation experiment part is completed.The experimental results show that this scheme can effectively realize unmanned Panoramic ring view of the ship.