The Research and Application of the Unmanned Aerial Vehicle Image Matching Based on Feature Point
|School||Xi'an University of Science and Technology|
|Course||Cartography and Geographic Information Systems|
|Keywords||Local invariant feature Image Matching colour-sift Large_SIFT UAV images|
Image matching technology is the core content of computer vision and digital image processing , is widely used in target recognition , 3D reconstruction , stereo matching , motion tracking , and other fields . UAV remote sensing system platform is not smooth , the camera attitude stability , these factors gave match brings difficulties , therefore , suitable for different degree of overlap , the big rotating angle images of the same name features automatic matching algorithm , is a low-altitude remote sensing image Can be applied to production practice premise . UAV remote sensing image target matching study and explore in-depth study of the matching algorithm based on feature points around the feature point extraction , based on the characteristics of the feature point description and feature matching three aspects of the research and experimentation . The article 's main research content and innovations are as follows : 1. Analysis of the importance of UAV remote sensing images and application status , summarizes the classification and study of the status quo of the image matching technique to analyze the main problems faced by the UAV remote sensing image matching technology . 2 . Study based on the local characteristics of the image gray information , respectively, from the rotational invariance , scale invariance , affine invariant introduced a variety of feature extraction operator conducted a comprehensive evaluation of the properties of each operator . Feature descriptor system research and analysis . Proposed three kinds of methods to remove false matches , respectively, based on the the RANSAC method the linear orthomorphisms transformation model , the correlation coefficient method and the main direction angle method , the effect of the three methods through experiments . Based on the color information of the image colour-SIFT feature matching algorithm , and to compare the effect of the SIFT algorithm , SURF algorithm and colour-SIFT algorithm for UAV image matching . The SIFT algorithm performance large format aerial imagery processing defects and deficiencies imaging the block Large_SIFT algorithm , the algorithm can be a good application in the UAV image matching process . 6 . Designed a set of the UAV images automatic aerial triangulation the measurement system GodWork, system of UAV remote sensing image matching part , used the previously mentioned large_sift algorithm , epipolar constraints , pyramid matching strategy . .