微小零件图像二值化处理的研究

Research on binaryzation processing of microcmponent image

采用图像处理与分析方法进行微小零件尺寸及形位误差检测时 ,所获得的微小零件灰度图像的噪声较大 ,对其进行二值化处理时需要根据图像的灰度值确定一最佳灰度阈值。如若确定的灰度阈值偏高或偏低将对图像分析的精度产生严重的影响。为此本文提出确定灰度图像二值化最佳灰度阈值的两种方法 ,即“双峰法”和“信噪比统计法” 。

As the binaryzation processing is made,an optimum grey threshold must be determined according to the grey value of the image.If the fixed grey threshold is a little higher or lower,the accuracy of image analysis will be affected seriously.So this paper puts forward two methods to define the optimum grey threshold of the grey image binaryzation “two peak” method and “s/n statistics” method.The practice and limitation of these methods for measuring the mini part size or its form and position deviation have been demonstrated.