Discussion And Research Of Least Square Image Matching About Digital Photogrammetry

Image matching as an important technology in the field of image Processing,is animportant issue in the field of Photogrammetry and computer vision.It is widely used inresource analysis,medical image registration,portrait and fingerprint recognition,radartracking and monitoring,aircraft cruise guidance and other fields.Nowadays,image matchingis used more and more,its accuracy can not meet the requirements of a lot of work,LeastSquares Matching attracted widespread attention because of its high accuracy.The paper use the correlation coefficient algorithm to complete the lea st squares imagematching, by studying the algorithms of generation of epipolar image,point fea tureextraction,image matching and least squares image matching, in the case of the Least SquaresMatching modules involved in the research and development of te rrestrial photogrammetrysystem. The image can reach sub-pixel level.In this paper,according to the characteristics of the original image,select the epipolarimage generation method based on digital rectification.In order to select the proper pointfeature extraction algorithm,this paper tests the Harris operator,Moravec operator and Forstneroperator,Analyses the pros and cons of each operator,Harris operator is selected as the datasource of the image matching.In addition,when the paper use the correlation coefficientalgorithm for image matching and least squares matching,analyses the effect of imagetexture,the search window size and the selection of threshold,try the best to make the imagematching with high efficiency,better matching results.