产品表面图像中的缺陷自动检测方法研究

仔细设计产品表面缺陷检测的图像识别算法是基于图像处理的产品表面缺陷自动检测系统快速并正确地检测缺陷的关键。以E型磁环表面缺陷的自动检测为例,分析了产品表面图像的特征,用梯度直方图自适应阈值分割的方法对图像进行增强和分割;并提出了一种新的快速识别算法对分割后的目标图像进行检测,实验结果表明了该算法的有效性和实用性。

Research on method of auto-detecting products surface image's defects

To design recognition algorithm Carefully is key factor that product surface defects auto- detecting system works perfectly and accurately. Taking example for auto- detecting surface defects of magnetic tach of E- shape, characters of surface Image of product are analyzed; Image is enhanced and segmented synchronously using the method of auto- finding threshold based on grads- histogram; thus a novel algorithm is presented to process the target Image. The Results of experiments show that the new algorithm is effective and practicable.