Aerial Images Matching Study Based On SIFT Algorithm

Image feature points extraction and matching is an important step in digitalimage processing, which is widely applied in technologies such as image mosaic,image automatic relative orientation, object recognition, etc.This paper adopts SIFT Algorithm on experiments and tests of aerialphotography image automatic matching. The results shows that the conventionalimages and problem image are well matched by the application of SIFT Algorithm.Due to the fact that SIFT algorithm is low and volumes of the feature points arelarge, this paper deducts the size of the image during the feature points extractionphase to reduce the data volumes. Then it sets a threshold value to make the ratiovalue of the similarity sub-maximum value to the similarity maximum value is lessthan the threshold value. By setting the threshold value to control the feature pointsnumber, the matching accuracy is largely improved. Moreover, through the adoptionof block matching strategy and matching range setting method and deduction of useof the exhaustive matching, the efficiency of matching was substantially increased,which realizes the practical application of SIFT algorithm on aerial photographyimage processing.