Moving Target Classification Technique in Video Sequence

Abstract

A moving target classification method using Krawtchouk moment and K-means clustering is proposed in this paper in order to achieve effective classification of moving targets. By introducing Krawtchouk moment and K-means clustering, the method can classify the targets into four types and achieves accurate classification results by extrating the low-level Krawtchouk moment invariants of the target images as feature vectors to prevent the influences of the targets scale and posture changes and clustering the feature data obtained according to K-means clustering algorithm with proper parameter. At last, the experiments verify the effectiveness of the method proposed. And result shows that this method has effective classification result. In addition, compared with the methods using Hu moment invariants with K-means clustering and Zernike moment invariants with K-means clustering, the classification rate of the method proposed in this paper is better than the other two methods.