Aiming at the faults in the internal thread detection,such as, poor efficiency, damage caused by contact detec-tion and so on,a internal thread efficient measure method based on machine vision was proposed in the thesis. Firstly,according to the nut features to be measured,a method of image acquisition was designed,which can acquire high quality images of internal thread,and the high quality were expressed in clear thread profile,high contrast,fast speed acquisition. And,according to the features of the images acquired, contrast ratios of images were enhanced by taking a method that gray images selectively, and weakening effect of scattering from surface,secondary noise reduction was based on the area of the single connected domain after filtering,and mending cracks by using the close operation,eventually internal thread image detection issue can be turned into a relatively mature external thread image detection one. The precision for thread internal diameter detection can be up to 12.78 μm and average time is no more than 300 ms from image acquisition to the end of the parameter detection.%针对目前内螺纹检测方法中存在的检测效率低,对工件接触性检测会造成损伤等问题,提出了一种基于机器视觉的内螺纹高效检测的实现方法.首先根据待测螺母的特点,设计了一套内螺纹的图像采集方法,利用该方法采集的内螺纹图像,螺纹牙型清晰,对比度高,采集速度快.之后针对所采集的内螺纹图像的特点,利用有选择性的图像灰度化方法增强图像对比度,同时降低金属散射对图像的影响,基于单连通区域面积进行二次降噪,利用闭运算连接螺纹图像轮廓断裂处,最终将内螺纹图像检测问题转化为相对成熟的外螺纹图像检测问题.对螺纹内径的测量精度达到12.78 μm,单个工件从图像采集到参数检测结束,平均用时不超过300 ms.
展开▼
