Matching of transmission lines monitoring photos

A major problem in image-based research studies on transmission lines monitoring is to find the corresponding points from different images of the same object. In the paper, matching the pictures of transmission lines which are photoed by the camera on the iron tower with different parameters are studied, we use sub-window registration, symmetric features and Harris corner detection to obtain fast and accurate rigid matching. First, the images are sub-blocked, and then the subimages are filtered using Log-Gabor filters to extract symmetric features. Sub sequently, Harris corner features are identified in each subimage and matched to corresponding points in the other subimages, and the best matching point pairs are selected. Finally, sub-images around these two sets of points are registered on the given basis using mutual information.