Research Of Image Rectification Based On S Factor In RTCam Camera Model

The research of binocular vision system has been a very important research field in computer vision.The research of linear camera's epipolar geometry has some results,but the study of nonlinear camera's epipolar geometry is still in the exploratory stage.The algorithm which solves the epipolar geometry of nonlinear camera based on image sequence solves the problem that the epipolar geometry of nonlinear camera is difficult to solve.RTCam camera model use a unique factor function which is in a high dimensional space incorporate a variety of popular camera models into a unified geometric framework,so we can use the RTCam camera model study a variety of popular camera models.We combine the solving epipolar geometry of nonlinear camera based on image sequence with RTCam camera model.We use the epipolar curve which is solved by the algorithm based on image sequence extract s factor information of RTCam camera model and use the s factor correct the distorted images collected by the wide angle lens.Firstly,this paper describes the principle of RTCam camera model in detail.RTCam camera model is essentially a projection conversion and RTCam camera model incorporate different linear and non-linear camera models into a unified geometric framework.This paper describes the significance of Q matrix parameters in the RTCam camera model in detail.We use Matlab simulate the 2D and 3D models of RTCam camera model and analyze the influence of s factor on the nonlinearity of image detailedly.Secondly,this paper realizes the algorithm which solves the epipolar geometry of nonlinear camera based on image sequence.We narrate the theoretical basis,model analysis,algorithm idea and smoothing constraint of the algorithm which solves the epipolar geometry of nonlinear camera based on image sequence.Finally,we analyze the robustness of the algorithm which solves the epipolar geometry of nonlinear camera based on image sequence,in the case where the image has Gaussian noise,Gaussian blur or image is out of focus.The algorithm solve the the epipolar geometry of nonlinear camera successfully.We use the traditional linear algorithm solve the nonlinear geometry of the camera model(wide-angle camera model)and analyze the results.We analyze the precision of the algorithm which solves epipolar geometry of the nonlinear camera model based on the image sequence.We analyze the precision that we use traditional linear method solve the epipolar geometry of nonlinear camera model.We extract the s factor in RTCam camera model from epipolar curve which is solved by the image sequence algorithm and use the s factor to correct the distortion image acquired by the wide angle camera.We analyze the precision of the s factor in RTCam camera model.