Research On Fixed Point Matching Technology Of Footprint Image And Linear Array Image

With the rapid development of remote sensing technology,satellites equipped with line sensors,area array cameras and laser ranging to achieve medium and large scale mapping has become the mainstream technical solution.In the subsequent linear array image bundle adjustment,the image point coordinates and laser ranging information or its conversion information are the main adjustment observations,and the image coordinates are matched to the linear array image(homologous)to obtain the connection point.There are also image point coordinates contributed by the laser footprint.So far,the matching technique between homologous images has been mature,and the matching between heterogeneous images is still a hot topic of research.In this thesis,the**satellite parameters and laser ranging data are used for simulation.The constellation of the laser footprint on the footprint image is used as the reference point.A series of matching schemes are developed to capture the image on the footprint.The feature of the reference point is used as the matching entity,and the point of the same name is found on the linear array image,that is,the fixed point matching.Therefore,this thesis studies the fixed-point matching technology between the footprint image and the linear array image,which lays an important foundation for the bundle adjustment of linear array image.The research content and innovations of this thesis are as follows:(1)In order to avoid fixed-point matching,pixel-by-pixel search on the linear array image,this thesis estimates the a homography matrix H as a priori knowledge,which provides convenience for narrowing the range of fixed-point matching.In addition,this thesis improves the algorithm for estimating the H matrix,making its algorithm relatively stable.(2)Develop a fixed-point matching scheme,and propose fixed-point matching based on Daisy operator,based on correlation coefficient and pyramid correlation coefficient,fixed-point matching based on single-play phase,fixed-point matching based on distance weighting,and verify each algorithm through experiments.And put forward a "Rematching"strategy to quantitatively analyze the accuracy of each algorithm,and finally determine the best fixed-point matching scheme.