关键词: 自动聚焦, 聚焦窗口自适应选择, 人工鱼群, 小波分析

Abstract: Window selection is an important module of autofocusing system（AF），which directly impacts the sharpness and accuracy of focusing evaluation curve and the complexity of AF system.It can reduce the processing data and increase the focus speed;it can also weaken the impact of non-interest region and improve the focus accuracy.For the traditional window selection algorithms have some limitations，an algorithm based on artificial fish swarm algorithm（AFSA） is proposed;because the standard AFSA falls into local optimum easily，the parameters’ setting and behavior are improved in this paper.Experiment results show that foreground image can be segmented accurately using this method.The proposed method can track the main object and be applicable to images with small contrast，so the window is dynamic with good adaptability.

Key words: autofocusing, adaptive selection of focus window, artificial fish swarm algorithm, wavelet analysis