Intelligent recognition technology of high speed moving target contour in dynamic visual scene

High speed moving targets can cause rapid distance changes between them and visual sensors such as cameras, and even lead to the phenomenon of crossing distance doors, where the target moves from one detection area to another in a very short period of time. If the target is accompanied by high acceleration, motion blur will occur during the imaging process, increasing the difficulty of contour extraction and causing deviation in target recognition. Therefore, research on intelligent recognition technology for high-speed moving target contours in dynamic visual scenes is proposed. Optimize the brightness of continuous dynamic visual scene images and, based on the optimization results, use a joint target area extraction algorithm to extract high-speed moving target areas; based on the setting of snake curves in the motion target area, a gradient vector flow snake model is constructed to complete the extraction of the edge contour of the motion target. On the basis of extracting edge contours, the contour description method based on the increment of centroid height is used to extract the feature of contour centroid height increment, which not only considers the problem of motion blur but also incorporates the dynamic information of contour changes with target motion. The joint dynamic programming algorithm is used to analyze the similarity between shape features, achieve target matching recognition, and reduce target recognition bias. The experimental results show that when using this method for high-speed moving target recognition, the recognition effect is good and the accuracy is high.