Research On The Treatment Of Void Halo In Image Mosaic

Due to the limitations of existing imaging equipment, we can not simply rely on the camera or video camera to get wide-field, high resolution images, therefore image mosaic technology came into being. Image mosaic technology is an important part of research in areas such as computer vision. With the rapid development of science and technology, its application involves all areas of social life, such as medical, military, geographic mapping and so on. It has become a hot topic in the field of image processing, and has a wide range of engineering applications.Firstly, this paper focus on the status of image mosaic technology research in detail, and analyzed problems in relation to the seam and uneven illumination. And for features of mosaic image we got, image scale, rotation, etc., we choose SIFT scale invariant feature transform algorithm. After that, we analyzed the ideas and concrete steps of the algorithm in detail.This paper analyzed vignetting of images produced in the phenomenon of non-ideal the camera hardware or uneven illumination conditions, and summarized causes and compensation method of vignetting. This paper proposed the method of based on Gaussian surface vignetting fix in the last. Take this add to the image fusion, and compensate exposure difference between the images together with fusion method of based on gradient pyramid. The experimental results show that this method reach a better transition effect in dealing with some images exist differences in exposure.Considering the image rotation angle and a large-scale changes, the effect of projection plane exits distortion phenomenon. This paper proposed the idea of cylindrical projection transformation of mosaic images, and the final panorama is processed in secondary cylindrical projection transformation. The experiment shows that the result of this method is more natural, so this idea has a higher value for eliminating image distortion.