摘要
介绍一种基于显微成像法的快速检测机加工零件表面质量的方法。通过对切削加工零件表面纹理图像进行灰度信息统计分析并与传统方法检测粗糙度结果做曲线拟合,建立加工表面轮廓高度特性与图像灰度特性之间的相关关系,从而根据图像灰度分析得到粗糙度值;用适当大小的窗口逐行扫描图像,通过检测窗口内波峰线的连续长度而识别纹理缺陷,并以窗口轮廓作为缺陷的标记。实验验证了该方法检测粗糙度等级准确度达83.3%,可识别并标记的纹理缺陷大小为30μm。
A method of rapidly detecting the quality of machined surface based on micro-imaging is introduced. The grayscale of machined surface＇ s texture image is statistically analyzed, and then compared with the roughness detected by traditional method for curve fitting. Therefore a relationship between the characteristics of contour height on the machined surface and the image grayscale is established. According to the relationship, roughness of machined surface can be obtained from the texture image. The texture image is scanned by using a window with appropriate scale, and the continuous length of the peak lines is measured to rec-ognize and mark the texture defects. The accuracy of roughness measured by the introduced method can reach 83.3% , and the di- mension accuracy of the defects can reach 30μm or more accurate.
出处
《现代制造工程》
CSCD
北大核心
2012年第11期91-96,共6页
Modern Manufacturing Engineering
