航空发动机叶片磨抛加工刀纹特征提取

Feature Extraction of Tool Pattern in Aero Engine Blade Grinding and Polishing

针对航空发动机叶片自动化磨抛加工缺少叶片表面快速检测技术的问题,以机器人磨抛加工为平台,搭建了视觉检测系统进行叶片图像采集。提出了基于灰度共生矩阵与加权小波分解的刀纹特征提取与识别方法,采用训练支持向量机分类器,并通过边缘检测算法提取刀纹边缘。实验结果表明,所提方法准确率为97.19%,可以精准地获取叶片刀纹大小和位置,具有准确、效率高、易于实现自动化等优点,可为航空发动机叶片磨抛加工自适应检测提供参考。

Aiming at the problem of the lack of rapid detection technology for the blade surface in the automatic grinding and polishing of aero engine blades,in this paper,based on the robot grinding and polishing processing platform,a visual inspection system is built to collect the blade images.We propose a method of feature extraction and recognition of tool pattern based on gray level co-occurrence matrix and weighted wavelet decomposition,use training support vector machine classifier and apply edge detection algorithm to extract the blade edge.The experimental results show that the accuracy of the proposed method is 97.19%,which can accurately obtain the blade blade size and position.It has the advantages of accuracy,high efficiency and easy to realize automation,which can provide reference for adaptive detection of aero-engine blade grinding and polishing.