基于MATLAB图像处理的直齿圆柱齿轮参数测量

运用MATLAB图像处理技术对直齿圆柱齿轮的参数测量方法进行了研究。通过机器视觉系统获取齿轮图像,并对其进行灰度化、去噪、二值化操作。按8邻域寻找齿轮孔区域并去除,运用Canny算子进行边缘检测;通过计算包含齿轮孔轮廓的最小矩形的尺寸得到齿轮孔直径。以8邻域边界跟踪法获得齿轮的单像素外轮廓。通过神经网络对外轮廓点分类的方法找到齿顶位置点和齿根位置点,运用最小二乘原理的圆拟合法计算齿顶圆和齿根圆直径;通过对齿顶或齿根图像进行膨胀操作后计算8连通区域的方法确定齿数;通过齿顶圆公式计算模数。编写齿轮参数测量软件并对6个直齿圆柱齿轮的参数进行了测量;通过与理论值的比较,验证了该方法的合理性。

Parameter Measurement of Straight Spur Gear Based on MATLAB Image Processing

The parameter measurement method of straight spur gear was studied by using MATLAB image processing technology. The gear image was obtained by using the machine vision system, and the image was grayed, denoised and binarized. The region of gear hole was found and removed according to 8 neighborhoods, and the edge detection was carried out by using the Canny operator. The diameter of the gear hole was obtained by calculating the minimum rectangular size containing the gear hole profile. The single-pixel outer profile of the gear was obtained by using 8 neighborhoods boundary tracking method. Position points of addendum and dedendum were found by using the classification method of outer profile points based on neural network. The diameters of the addendum circle and dedendum circle were calculated by using the circle fitting method based on the least square principle. The number of teeth was determined by using the method of calculating 8 connected regions after expansion operation of the addendum or dedendum image. The modulus was calculated by the addendum circle formula. Gear parameter measurement software was prepared and the parameters of 6 straight spur gears were measured. The rationality of the method was verified by comparing with the theoretical value