Research On Text Image Stitching Technology And Its Application

Image stitching is a hot topic in the field of computer vision.It is widely used in modern applications such as high-resolution reconstruction of satellite photos,medical image processing and video stitching.Image stitching is to splicing multiple images from different viewing angles into one image at the same viewing angle.Since the camera's shooting angle is different,the overlapping regions of the two images cannot be simply overlapped by image copying.The main purpose of image stitching is to solve the problem that the wide viewing angle cannot be obtained due to the limitation of the shooting device.The wide viewing angle image can be directly realized by a wide-angle lens or a normal lens plus a spherical lens,however,the hardware cost is high.The low-cost solution is an ordinary camera combined with an image stitching algorithm and an ordinary camera can perform a similar wide-angle lens effect by multiple shots,so more image processing methods applied to image stitching are proposed.In this paper,the image stitching technology based on text image is studied.Given a group of images with considerable overlap,it can be spliced into a wide-angle image.The main work of this paper is as follows:(1)The background meaning of image stitching technology and its research status at home and abroad are introduced.(2)Three key aspects of image stitching technology are studied:image preprocessing,image registration and image fusion.The image registration method based on SIFT is introduced and the common image fusion algorithm is introduced.(3)In the image preprocessing stage,the color information of the image is often affected by factors such as the color deviation of the light source and the acquisition device,which will reduce the accuracy of feature matching and negatively affect the image registration,and propose a fast automatic color enhancement algorithm.The experimental results show that the proposed algorithm is richer in the color information of the processed image and more image feature points can be found.(4)In the image fusion stage,the pixel correction and brightness adjustment of the overlapping regions are performed on the basis of the HIS space and the stitching seam due to the chromatic aberration can be effectively eliminated.Based on the above method,the realization process of image stitching is demonstrated through experiments and the experimental results of image stitching are given.The image stitching based on the method is faster,the image features more points and the registration effect is better.