Point-based Study Of Image Stitching

Image mosaic technology refers to the existence of overlapping image sequences from the same scene,different angles for image registration,and to fusion a high-definition images that contains all the image information.Image mosaic technology is the increasingly popular are of technology,and with the continuous advancement of computer technology has been widely used in space exploration,conservation,medical image processing,military reconnaissance,police forensics,digital video compression,and virtual reality technology and other fields.In computer vision,image mosaic technology has been very important and one of the studies are relatively complex.The whole process of image mosaic technique mainly consists of the image pre-processing,image registration and image fusion.The image registration problem is the key and core in image stitching technology.The focus of research image registration is how to improve the speed and accuracy of registration,and to improve the adaptability of registration algorithms.In general, the image registration method can be divided into gray-scale image registration,transform domain-based image registration and feature-based image registration.Gray-scale image registration and transform domain-based image registration calculation are too large,and can not well resolve the issues such as diversity and rotation angle,when we use the camera for image acquisition these problems are very common,so these two methods are not well applied in the actual operation. Feature-based image registration method,in particular,is based on the point of image mosaic technology has advantages of computation,high precision, adaptability,and information advantages of large,and also has good adaptability between the image noise,rotation and scale differences,through continuous research and development,based on point feature image mosaic method has been widely applied in computer graphics and image processing in many aspects. An image mosaic method based on SIFT and Selection Region is described in this paper.Firstly,it uses scale invariant feature transform (SIFT) to obtain the image feature points.Secondly,makes full use of the rotational invariance of circles and the maximum mutual information for feature points matching.In the end,in the problem of image fusion,use the weighted average method and the introduction of the concept of the threshold to achieve a smooth transition of images overlapping regions.In this paper, different photos selected simulation experiments proved the feasibility of this method.Experimental results demonstrate that the proposed method is robust to translation,rotate,intensity and noises,and can produce high quality image mosaic.