基于统计特征的微型轴承滚珠计数检测

Counting Detection of Miniature Ball Bearings Based on Statistical Characteristics

随着传统行业人力成本提高,传统轴承行业的检测技术逐渐向着将向高速化、网络化、信息化和系统化的方向发展。面向传统轴承产业的大规模的生产与检测,实现一种基于统计特征的微型轴承滚珠计数检测算法,通过主动搜索最佳阈值范围,实现图像的区域划分,并依据已知的轴承尺度信息,快速定位轴承中心,沿着轴承滚珠分布方向,统计其灰度值变化的规律,找到最佳的滚珠中心点,并统计滚珠数量,该方法的抗干扰能力很强,算法效率高。

Along with increasing human cost in traditional industry, bearing detection technique leads to high precision, high speed, networking and information system. Catering to bearing production and testing, implements an adaptive detection of miniature bearing ball based on statis- tical feature, realizes the image regional division by the active threshold searching, adaptively locates the bearing center by object size constraint, then finds the best ball center by calculating the statistical gray value along the ball distribution direction, counts the balls. The technique is robust and greatly improving system efficiency.