Font Size: a A A

SIFT Based Image Of Low Altitude Photogrammetry Matching Algorithm Research

Posted on:2017-08-05Degree:MasterType:Thesis
Country:ChinaCandidate:K N ZhangFull Text:PDF
GTID:2348330503474616Subject:Resources and Environment Remote Sensing
Abstract/Summary:PDF Full Text Request
In recent years, aerial photogrammetry has become an important method of mapping and updating geographic information database, and has developed rapidly in many fields. Low altitude photogrammetry has the advantages of strong operability, low cost, high speed and high precision, it has become a hot spot in the research of photogrammetry technology. Low level image matching technology in image fusion, 3D modeling, object recognition etc.increasingly widely used, some factors give the matching difficult. Pure traditional image matching technology has not improved the low altitude photogrammetry image matching problems that exist in the process. Therefore, urgently need a perfect method can be for low-level image matching. The problem of image were research and discussion, combined with the existing image matching operator and presents a method for low-level image feature points extraction and matching. The main work and research results are as follows:1. Analyzes the necessity and the importance of feature point matching, with induction summarizes the pretreatment method and the process of image matching, analysis the key technology in the process of matching, made clear the purpose of this paper.2. This paper mainly studies SIFT operator, analyzes the field of Photogrammetry in several commonly used feature point extraction Moravec, SUSAN, Forstner and Harris operator, and through the experiment, the operator speed, accuracy, limitations and adapt to the comprehensive evaluation, to provide basis for research on the algorithm.3. According to the advantages and disadvantages of SIFT and the above four kinds of operators, this paper proposes a method which is suitable for photogrammetry image.Experimental show that: simplified feature descriptors and Gaussian scale space can improve the image matching efficiency, using Harris operator and Forstner operator precise positioning of feature points can greatly improve the matching accuracy, the method can be widely suitable for application of low altitude photogrammetry image, the matching results compared with other algorithm has more reliability and stability, has a certain practical significance.
Keywords/Search Tags:low altitude photogrammetry, image matching, SIFT, Harris, feature point detector
PDF Full Text Request
Related items