Gabor滤波器在自动视觉表面检测中的应用  

Application of Gabor Filter in Auto-visual Surface Detection

作　　者：温兆麟[1] 陈新[1] 李克天[1] 郑德涛[1] 

机构地区：[1]广东工业大学

出　　处：《中国机械工程》2007年第15期1838-1841,共4页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(50475044);教育部科技研究重点资助项目(2004106);广东省科技攻关项目(2005A10403004);广东省自然科学基金资助项目(7007362)

摘　　要：提出一种基于Gabor滤波器的纹理表面缺陷检测方法。该方法基于Gabor滤波方案,计算纹理图像与某一Gabor滤波器卷积能量输出响应。选择Gabor滤波器的最优参数,采用简单的二值化技术将纹理差别转化为可检测的滤波器输出。纺织品、铣磨面等结构纹理和皮革、砂纸等随机纹理的实验结果证明了该方法在识别不可定量测量缺陷方面的有效性。This paper put forward a series of defect-detecting methods of texture surface based on Gabor filter. Utilizing Gabor filters, the method calculated the energy-output response between texture images and the convolution of a certain Gabor filter. With selection of the best parameters and application of a simple technique of dual-value, texture discriminations were transferred into detectable output of filters. The method was tested on a number of structural textures of textiles and milled surfaces, as well as stochastic textures of leather and sand paper, the outcomes justify the method＇s effectiveness in terms of immensurable defect-detecting.

关 键 词：机器视觉 表面检测 缺陷检测 GABOR滤波器 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]