关键词: 目标检测, 小波域, 聚类分割, 最佳阈值

Abstract: Because of the contradiction between accuracy and real-time in monitoring system, the object detecting system in wavelet domain based on clustering segmentation algorithm is researched. A method to determine the adaptive threshold for high and low frequency components is presented. The judging criterion and the optimum threshold are deduced. The algorithm can remove the system noise, detect the object fast, and ensure the precision of system. It is proved by simulation and experiment results that the algorithm can detect the object accurately and realize a real-time system.

Key words: objects detection, wavelet domain, clustering segmentation, optimum threshold