Skip to Main Content
In order to deal with the problem of that conventional algorithm of mosaic needed large coincidence region for registration and different gray in coincidence region may influent quality; this paper presents an algorithm of mosaic that combing the Fourier - Mellin Transform (FMT) and the characteristics of the Regional Detection. Firstly, the feature images are intercepted from real images for the information of shift, rotation and zoom by FMT. Then combines the location information that the feature images in panoramic image to splice image. Finally, we can mosaic image correctly in the condition of release workload. The experimental results show that algorithm can decrease the region overlay in adjacent image, on condition that ensure the stitching precision to improve splicing efficiency, while the Conventional image mosaic algorithm likes Fourier - Mellin Transform (FMT), needed above 50% coincidence region in adjacent image to realize correct mosaic.