Font Size: a A A

Research On The Application Of Stereo Marching Technology In The Photogrammetry Of Wave Images

Posted on:2017-05-06Degree:MasterType:Thesis
Country:ChinaCandidate:F M YinFull Text:PDF
GTID:2322330518972938Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
In the application background of measuring waves by binocular photography technology,for the purpose of increasing the number of matching pairs and reducing the error matching rate, the paper studies on the feature extraction algorithm and feature matching algorithm of stereo matching technology of binocular vision, and utilizes the improved stereo matching algorithm in the waves measurement.Firstly,classical feature extraction algorithm Moravec?Harris and SIFT are introduced,and through simulating and extracting feature points of wave images, the experimental results show that the feature points of Harris algorithm can reveal the texture feature of waves most clearly and accurately, and the SIFT feature extraction algorithm has the best stability when the wave image changes variously.Next,the pyramid matching algorithm and the SIFT matching algorithm are introduced,and through simulating and matching wave images, the experimental results show that the error matching rate of SIFT matching algorithm is not high in the application of wave images.And then, for many shortcomings of SIFT algorithm, such as the number of matching pair of wave images is too little, and error matching, this paper introduces the improved stereo matching algorithm to process wave images. In the feature extraction section, Harris sub-pixel corners substitute the DoG extreme points of SIFT,so that changing the threshold of Harris corner response can change the number of comer. Sub-pixel coordinates of Harris corners are obtained by approaching to the corner response value of the integer coordinates by the quadratic polynomial. In the feature matching section, using the two-direction matching in SIFT matching to eliminate mismatch pairs, and using the epipolar constraint to determine the candidate matching set in the two-direction matching strategy. Compared with traditional SIFT algorithm, experiments show that not only the error matching rate of improved algorithm is lower, but also the number of matching pairs has increased obviously.Finally, in order to verify the improved stereo matching algorithm on the applicability of the wave images,processing wave images by the improved stereo matching algorithm in the Laboratory Tests and the Circulating Water Channel Tests, and matching results are used for drawing up 3D map and contour map in the three-dimensional reconstruction. Experiments show that the waveform and wave direction of 3D map are the same with actual waves, and the errors of the measurement wave height and the actual wave height are about 2.67% and 4.84% respectively of the two sets of data results at the mark in the Circulating Water Channel Tests.
Keywords/Search Tags:waves measurement by Binocular photography, SIFT algorithm, sub-pixel corners of Harris, two-direction matching, the epipolar constraint
PDF Full Text Request
Related items