一种改进Canny算子在织物疵点检测中的应用

Detection of Fabric Defects Based on an Improved Canny Operator

针对经典Canny算子高、低阈值的选择存在被动性,实时性差,不适于在线检测的问题,给出了一种改进的Canny算子。该方法应用最大类间方差法自适应地确定Canny算子的高、低阈值,以提高Canny算子边缘检测的实时性和准确度。实验结果表明,这种方法不仅改善了Canny算子的实时性,而且具有更好的抗噪声能力和检测准确度,提高了织物疵点在线检测的实用效果。

Being aimed at solving the classical Canny Operator＇s problems that it is passive to choose highthreshold low-threshold and its bad real-time capability makes it not be suitable for on-line monitoring,this paper presents an improved Canny Operator.In order to improve the real-time level and accuracy of edge detection based on Canny Operator,the maximal variance between-class has been used to determine Canny Operator＇s double threshold adaptively.Experimental results have shown that the method not only improves Canny Operator＇s real-time performance,but also has better anti-noise capability detect accuracy,so it improves the practical function of on-line detecting fabric defect.