Research on the Moving Vehicle Detection Algorithm Based on the Motion Vector

摘要

In this paper, moving object detection under the dynamic background in aerial videos has been studied, and a moving vehicle detection algorithm based on the motion vector is proposed, using the KLT algorithm for matching feature points, according to K-medoids clustering algorithm on the feature points, finally the location of the target vehicle has been found. Experiments show that the method in this article can accurately detect the moving vehicles in the aerial videos, and then extract the corresponding traffic information.