Video Splicing Based On Feature Detection Of Modified SURF Algorithm

The video splicing is an extended technology based on image mosaic. It has a broad application prospects in many areas such as video surveillance, television and other fields. Now there are many ways to get a panoramic image, for example, using an ultra wide-angle lens, an professional camera with panoramic lens is also be a good equipment, fish eyes cameras and so on. But there also exist problems of expensive, a bad flexibility and a limited rotation angel. This paper presents a new method for image mosaic. Detecting video frame using improved SURF algorithm, then completing the feature points matching process, finally design an real-time algorithm for video-frame mosaic.First, this paper describes and introduces the process of image registration and method of image registration in detail, while presenting some methods of image fusion. This paper gets an idea through image matching; image fusion based detecting feature points to get a panoramic image.Second, this paper also analysis the classical algorithm for feature detection. Such as SIFT algorithm and SUFR algorithm, and presents a modified SURF algorithm which limits the area detected. In this way, image mosaic can find a transformation matrix in an overlapped area where feature points will be found. When detecting feature points, we can set a detection radius to limit the number of feature points detected. The concrete process described as follows:choose an point in the set of feature points, then set this points as the center of an circle in which only one points will be selected and other will be removed.Finally, we use RANSAC algorithm to remove the false matching pairs, then get transformation matrix through Least squares method. Fuse the image found by matrixing and gradually weighted average method to form an panoramic video. In the end of this paper, a design proposal is presented for auxiliary vehicle-video sy stem based on video mosaic scheme which proposed by this research.