The Research Of Stereovision System Based On Fisheye Lens

In recent years,with the continuous development of artificial intelligence,computer vision technology has been more and more widely used.The needs of many occasions have shown the limitations of the scope of ordinary lens perspective.The fisheye lens has a large field of view,you can get a one-time hemisphere or even 360 ° airspace scene information,far more than the ordinary lens vision system' shooting field,but also greatly beyond the human natural observation capabilities,this advantage making it attracte more and more attention in the civilian,industrial,military areas.Among them,fisheye lens binocular stereoscopic vision technology is one of the hotspots in computer vision research,and has great prospects in large-scale three-dimensional reconstruction and robot positioning and navigation.In this paper,the calibration of the binocular fisheye camera in the air and the calibration of the underwater binocular fisheye camera,the stereo matching of the fisheye images in the air and the stereo matching of the underwater fish-eye images are systematically studied.The main research work is as follows:Firstly,a calibration method of fisheye camera based on centroid optimization is proposed.The basic template calibration method is studied.In the process of calibration,there is a problem that the center of the calibration plate is not coincident with the elliptical centroid of the image,therefore,the particle is optimized.The center of the ellipse and the elliptical centroid of the image are calculated and the new objective function is established according to the calculation result.Then the camera calibration parameters are optimized and the accuracy and stability of the method are systematically analyzed.Secondly,a stereo matching algorithm based on polar curve geometry and variable support window is proposed.Considering the problem of the distortion of the fisheye lens,the polar curve equation of the system is deduced according to the optimal projection model of the fisheye lens,and the search range of the corresponding point is determined by using the polar curve equation.According to the the homologous points' characteristicof the position relation on the left and right images,the support neighborhood of each center pixel point is determined and the corresponding support neighborhood of the support neighborhood in another image under different disparities is calculated.And using the neighborhood calculates the matching cost of each point.Thirdly,the refraction model of underwater binocular fisheye camera system is established,and a two-step calibration method of underwater camera is proposed.The basic imaging parameters of the camera are calibrated by Zhang Zhengyou's checkerboard plane calibration method.Then,the parameters of the system are calibrated.the reference point of the underwater checkerboard with known world coordinates is projected onto the imaging plane.The coordinates of the image points are obtained by the modeling model,and the evaluation function is obtained by minimizing the actual image point coordinate and the model of image point coordinate.The calibration results are obtained by the particle swarm optimization algorithm and the underwater space point localization algorithm is studied.Finally,the polar curve of underwater image is deduced according to the underwater binocular fisheye camera system model.In order to make the subsequent matching algorithm more accurate,an adaptive window based on color segmentation is proposed and the whole stereo matching of the underwater fish-eye images is completed.The experimental results verify the correctness of the model and the derived polar line,and the constructed adaptive window can be better applied to the stereo matching algorithm of the underwater fish eye image.