基于模糊推理的聚焦曲线峰值搜索算法研究  

Research on Focus Curve Search Algorithms Based on Fuzzy Decision

作　　者：陈东[1] 杨建中[1] 王金[1] 

机构地区：[1]华中科技大学机械科学与工程学院,湖北武汉430074

出　　处：《电视技术》2014年第7期33-36,共4页Video Engineering

基　　金：国家科技重大专项(2012ZX04001-012;2012ZX04001-022);国家高新技术研究发展863计划项目(2009AA045004)

摘　　要：为了提高变焦成像系统中自动聚焦的速度与精度,提出了一种使用Zoom-Focus曲线和模糊推理的聚焦曲线峰值搜索算法。在峰值搜索过程中,通过Zoom-Focus曲线将搜索范围缩小至全程的50%以内;采用模糊推理的方法自适应调整搜索步长,在远离焦点时大步长快速搜索,在接近焦点时小步长准确定焦。实验结果表明,该方法可以将聚焦精度提升至5μm,在小倍率情况下将平均搜索时间缩短为0.7 s,满足自动聚焦系统快速、准确、鲁棒性强等要求。In order to improve the speed and accuracy of the auto-focus process in a zoom imaging system, a novel peak search algorithm based on Zoom-Focus curve and fuzzy decision is proposed. By adopting this new method, the search range is narrowed down to less than 50% of the whole range, and search steps can be adaptive according to the current focus value and its variance. When away from the in-focus position,large step is employed to approach the target position as quick as possible ,and then small step is adopted to locate at it accurately. The experimental results show that by using this method focus accuracy is promoted to 5 p^m,and average search time is reduced to 0.7 s in the wide end of zoom lens, satisfying the requirements of fast- ness,accuracy and high robustness of the auto-focus system.

关 键 词：自动聚焦 峰值搜索 模糊推理 Zoom-Focus曲线 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]