基于小波图像评价的复合自动对焦系统研究  

Research on composite autofocus system based on wavelet image evaluation

作　　者：李天剑[1] 王会香[1] 陈松[1] 张伯辉[1] 刘相权[1] 

机构地区：[1]北京信息科技大学智能机器人技术研究所,北京100192

出　　处：《北京信息科技大学学报（自然科学版）》2013年第5期38-42,共5页Journal of Beijing Information Science and Technology University

基　　金：2012年北京市教委面上项目(SQKM201211232004);北京市教委学科与研究生教育项目(PXM2012-014224-000040;PXM2012014224-0000422)

摘　　要：针对传统管道检测系统(PDS)对焦时延长、图像质量不佳及对焦位置控制误差大等问题,提出了一种由音圈电机驱动对焦的管道视频检测系统。视频检测系统采用自动对焦技术,图像评价算法选用基于小波变换和方差图像清晰度评价函数(SEF)的二阶复合算法,电机搜索策略选用改进的爬山法搜索策略。选用一组离焦—聚焦—对焦的模拟管道图片对所选图像评价算法在Matlab上进行验证,结果表明选用的二阶算法能实现优缺点互补。再将图像评价算法移植到视频检测系统进行实验验证,取得了理想的对焦效果。To solve the problem of long time delay of focusing, fuzzy image and motor control error in the traditional pipeline detection system （ PDS）, a kind of PDS with focus driving by voice coil motor is presented. The autofocus technology is used in the PDS. A two-stage composite algorithm based on the deviations function and the wavelet transform and the improved mountain climbing search strategy areapplied in the PDS. By using a series of defocus-focus-defocus simulated pipeline images to validate the selected image evaluation algorithm in Matlab, it is proved that the two order algorithm can achieve reciprocal advantages and disadvantages video detection system and achieves ideal Finally, the image evaluation algorithm is transplanted into the focusing effect.

关 键 词：管道机器人 小波变换 自动对焦 音圈电机 Matlab 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]