Study On Extracting Image Feature Information Algorithm

With the rapid development of computer vision and computer graphics, based on imagefeature information, exploring the imaging object has become a hot topic and has been alsoimportant role in the development of stereo vision. Therefore, describing the characteristicalinformation points of the target based on the angle of the target image has been extractedthrough edge detection of the image, feature point extraction, and feature points matching.Considering effect from factors like environmental noise, in order to be able to obtainthe appropriate description information of the target object image, the paper proposedimproved algorithm of image edge detection, the algorithm can extract accurate the imageedge for the fuzzy image. Respectively, Harris algorithm and SIFT algorithm extract thefeature points in the image information, according to different description information ofimage feature points: the traditional gray correlation method under gray information andEuclidean distance matching of the descriptor information will be considered to producelarger defects of mismatch. Therefore, the improvement different feature point descriptioninformation of matching algorithm is put forward, more and correct matching points is able tobe get. Based on the research of the theory of Stereopsis, on the basis of camera imagingprinciple, two eliminate false matching algorithm research: epipolar geometry method and theleast square median method, after comparing, the least-square-median method is adopted tofurther exact match for obtained matching feature points, after eliminating the false matchingpoints, the number of the points of mismatch points is greatly reduced. Finally, the features inthe solid space information point is extracted by the characteristics of the image Information.Simulation experiments for image feature information extraction algorithm is completedby using Matlab tools, through the experiment, large quantities of describing the imageinformation of feature points in a reliable way are obtained in the experimental results.Thus,the feasibility of proposed method is confirmed.