The Research Of Image Mosaic Technology Based On Weighted Poisson Fusion

The purpose of Image stitching is the formation of a high-resolution, wide-field image by operating multiple images with similar scene, overlapping portions through a series of matching, fusion algorithm processing. With the development of computer graphics and vision technology, image stitching technology is continually being innovative and effective solution to the contradiction between high resolution and wide field of view, it is widely used in medical image processing, virtual reality treatment, panoramic video monitor and other fields, But the technology exists defects for exposure differences sensitive and can not be a good solution to splice, it is still an important study topic in the area of image stitching processing.To solve these problems,our paper propose the interaction of Weighted Fusion and Poisson Fusion.Core algorithm includes a feature point extraction, feature matching and image fusion of three key modules, each module in turn calls. For consuming too much problem in feature point extraction module, scale spatial structure improvements and optimized code generation characterization process are used to further enhance the feature point extraction efficiency. Feature matching section uses a combination of coarse and fine matching policy, on the premise for ensuring the accuracy of the feature point set for further screening of the feature point filtering, the efficiency and accuracy of spatial transformation model have been improved, the convergence effect has also been improved accordingly. In order to achieve seamless integration, the algorithm uses the interaction of weighted average image fusion and Poisson image fusion, after the weighted average smoothing, the next step is the full image brightness uniform average treatment to reduce the exposure difference, and finally re-embedded coincide scene through Poisson fusion, representing a better solution to the problem of image stitching marks.By contrast with the conventional image stitching algorithm, our algorithm's time efficiency has improved, the ultimate fusion also has a greater difference. Experiments show that the improved algorithm thought feasible and strong application for the development of image stitching technology has a certain role.