Research And Implementation On Dynamic Scene Stitching

Image mosaic has been a research hot spot of the computer vision technology and image processing. It represents a high-resolution and smooth image with large view, which consists of a number of discrete images information, which well resolved the conflicting issues field of image view and resolution. In virtual reality technology, tourism, medical image processing, remote sensing measurement, aviation and military fields, the technology has a wide range of applications.The image mosaic can be divided to the static scene stitching and dynamic scene stitching. for static scene stitching, there are many algorithm which has reached perfect results. But for dynamic scene stitching, the existing algorithms are unable to overcome the challenges introduced due to simultaneous presence of Large exposure difference and Moving Objects.Firstly, this paper gives the process of the image mosaic, and completely summarizes the significance, the situation, the main application fields of image mosaic and the existing image mosaic technologies. The different kinds of algorithm of each key steps are analyzed and compared. Secondly, this paper introduces the detailed implementation of the dynamic scene registration algorithm based on SIFT characteristics, and simulate the dynamic scene registration algorithm. It gives the registration rendering of the dynamic scene at last.This paper introduces the problems existing in dynamic scene mosaics and proposes an improved algorithm about the optimal seam-line searching. I improve the optimal seam-line search rule with the gray sent to instead of color intensity difference to construct the optimal seam-line search rule. After constructed the optimal seam-line searching rule ,we search an optimal seam-line with dynamic programming and realize the image fusion with the poisson fusion. Experimental results show that the new algorithm can avoid the dynamic elements and realize the dynamic scene stitching. Due to the using of grayscale images in the process, reduces the complexity of the algorithm.