材料缺陷视觉检测及公差测量关键技术研究.pdf

年 月 2009 4 航空精密制造技术 Apr． 2009 第 卷 第 期 45 2 AVIATION PRECISION MANUFACTURING TECHNOLOGY Vol． 45 No． 2 材料缺陷视觉检测及公差测量关键技术研究* 刘琼，秦现生，蔡勇，宫养飞 （西北工业大学机电学院西安710072 ） 摘 要 在研究分析了竹板基材外形特点、表面缺陷特征的基础上，设计了竹板基材缺陷视觉检测及公差 [ ] 测量系统；根据双目视觉原理，经曲线曲率特征匹配方法，可计算得到光条上曲率绝对值最大值点的空间位 置和深度；结合基材尺寸的公差测量方法，可实现基材宽度与厚度的测量；针对竹板基材纹理表面的复杂性， 提出了基于小波分解与聚类思想的基材缺陷检测算法，实现了对基材缺陷的视觉检测；并用实验进行了验 证。 关键词 视觉检测；竹板基材；表面缺陷；公差测量 [ ] 中图分类号 ， 文献标识码 文章编号 （ ） [ ]TP391 TB9 [ ]A [ ]1003-5451 2009 02-0032-06 Research of Key Technology on Tolerance Measurement for Foundation Material by Computer Vision LIU Qiong, QIN Xian-sheng, CAI Yong, et al （College of Mechatronics, Northwestern Polytechnical University, Xian 710072 ） [Abstract] In order to achieve automatic defects detection and tolerance measurement for foundation materials of bamboo board, a high-performance, non-contact visual detection method was presented. Based on an analysis of the features of bamboo material shape and surface defects, the defects detection and tolerance measurement system for foundation materials of bamboo board using computer vision was designed. By means of curvature feature matching methods, the position and depth of the points on the laser strips could be obtained according to the principle of binocular vision. By means of tolerance measurement methods, the size of foundation material of bamboo board could be acquired. Owing to the complexity of colours and textures for bamboo, a defects detection algorithm based on wavelet decomposition and the clustering was presented. The key techn