基于图像的异型墙地砖在线分道系统  

Online classifying system based on image for shaped floor ceramic tiles

作　　者：刘鲤扬[1] 项基[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027

出　　处：《传感器与微系统》2012年第6期108-110,113,共4页Transducer and Microsystem Technologies

基　　金：国家"863"计划资助项目(2008AA042602)

摘　　要：提出了一种异型墙地砖在线提取和纹理比较的方法。将CCD采集的墙地砖图像进行灰度阈值分割后,经过以矩形模式为特征的轮廓搜索,对原始图像中的砖面区域做标准化仿射变换处理,实现墙地砖的在线提取。依据计算共生矩阵特征向量,采用最小距离分类原理进行纹理比较分类决策,并采用自适应阈值参数加强算法的鲁棒性。基于所提取算法研制了异型墙地砖在线分道系统,并在实际现场运行成功,取得良好的分道效果。An Acquiring and texture comparing method of shaped floor ceramic tiles is proposed,After the image of tile collected by CCD has been segmented by gray threshold,the search of contours judged by feature of rectangle patterns is used to acquire the normalizing affine transformed tile region.Based on co-occurrence matrix feature vectors,the minimum distance classifying principle is adopted to carry out texture classifying decision,and adaptive threshold parameter is implemented to improve the robustness of algorithm.Shaped floor ceramic tiles system run successfully in actual scene and show the classifying is effective.

关 键 词：墙地砖 图像分割 共生矩阵 自适应参数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]