A Method Of Linear Feature Matching For Aerial Images Based On Triangulation And MSLD Descriptor

With the large number of aerial digital images and satellite remote sensing images being gained,digital photogrammetry is widely used in obtaining attribute information of 3D spatial objects from images.Because of such large number of applications of digital image and 3D information,more and more attention has been paid to the high efficiency,automation and reliability of digital photogrammetry.As an important part in the field of digital photogrammetry,image matching is a research hotspot of computer vision and digital image processing and has become a common academic problem of world scholars.Compared to point feature,linear feature in the image contains more rich gray,texture and other information;Therefore,the line feature becomes an important matching element in image matching.This paper focuses on linear feature matching and presents a method of linear feature matching for aerial images based on triangulation and MSLD descriptor.Starting from the straight line extraction,this paper selects" LSD straight line extraction algorithm to extract straight line information in image after the experiment of three classic line extraction algorithms.Then,the epipolar constraint and the MSLD descriptor are used to match the line and good matching results are obtained.The main contents of this paper are as follows.In terms of line extraction,This paper analyzes and summarizes the existing algorithms of line extraction and introduces three methods of Hough transform line extraction algorithm,Freeman chain code line extraction algorithm and LSD line extraction algorithm.The experimental results of the three algorithms are compared and analyzed,and the advantages and disadvantages of the algorithm are summed up.Finally,the LSD line extraction algorithm is selected as the straight line extraction method,which has the advantages of fast running speed,less time consuming and high precision,and it is not easy to break.In terms of line matching,This paper presents a new method for linear feature matching of aerial images based on triangulation and MSLD descriptor.On the basis of obtaining the same points and LSD line extraction results in SIFT matching,the search range of the constraint line of the same triangle with the same name is firstly used and the overlap between the line and the candidate line is determined by the epipolar constraint,and thus establish a straight line support area.At last,according to the linear descriptor of MSLD,the straight line is determined.In terms of experiment analysis,The aerial digital images captured by different sensors are tested and the results are analyzed.The experimental results show that this method has high matching accuracy,and can obtain reliable results.