关键词: 图像拼接, SURF算法, 运动目标检测, 背景差分法

Abstract: The moving object detection of dynamic image, in the modern world is a key point, a difficult point and also a hotspot in image processing, it has a lot of useful application, such as keep a lookout over traffic situation in intelligent transportation, the precise guidance in military affairs and so on, but the recognition range of dynamic image has been the limitation of moving object detection. In order to solve this problem, this paper have made a research and put forward a kind of new method that using image mosaic technology to expand the scope of image recognition, and then based on the wide angle image we have get to finishing the detection of moving object. In this experiment, first we considered the problem of real-time and then we choose the SURF (speeded up robust transform) as an image matching algorithm instead of the SIFT (scale invariant feature transform) which is usually adopted in image mosaicking, because of the SURF algorithm can provide enough characteristic points and on other hand it’s working speed is faster than SIFT algorithm, and after the image mosaic we start to do the detection work about the moving object, at this time we took advantage of the background difference method, which can help us detect the moving object completely and quickly. Second, we considered the problem of practicability and then we choose the open source system