Abstract:

During industrial quality inspection, image segmentation of stamping defects is an important part of defect detection, which directly affects the effectiveness of defect detection. However, traditional Fuzzy C-Means (FCM) clustering algorithms overlook spatial neighborhood information and are sensitive to noise interference, resulting in poor segmentation accuracy. Furthermore, they are susceptible to the influence of initial values, which leads to a slower convergence speed. To address these issues, this study proposes an improved FCM algorithm, which replaces Euclidean distance with simple two terms of kernel-induced distance, and maps the original spatial pixels to the high-dimensional feature space to increase the linear separability probability and computation speed. The algorithm improves noise resistance and segmentation accuracy by utilizing the spatial correlation between image pixels and introducing an improved Markov random field to modify the FCM objective function. Using the Bald Eagle Search (BES) algorithm to determine the initial clustering center of FCM improves detection accuracy and convergence speed. Simultaneously, it also avoids the situation where the algorithm is prone to falling into local extremum. To validate the performance of the improved FCM algorithm, partition entropy, partition coefficient, Xie_Beni coefficient, and iteration number are used as evaluation indicators and compared with image segmentation algorithms proposed by different scholars in recent years through experiments. Experimental results show that the algorithm proposed in this paper has good noise resistance and can achieve good defect segmentation results, which implies a certain degree of application value for defect detection of stamping parts in the industry.

Key words: Fuzzy C-Means (FCM) clustering, industrial applications, stamping defects, kernel induced distance, Markov random field, Bald Eagle Search (BES) algorithm

摘要：

在工业质检过程中, 冲压件缺陷图像分割作为缺陷检测的重要环节, 直接影响缺陷检测效果。而传统的模糊C均值(FCM)聚类算法未考虑到空间邻域信息, 对于噪声干扰较为敏感, 导致分割精度较差, 且其整体易受初始值的影响, 造成收敛速度变慢。针对上述问题, 提出一种改进的FCM算法。采用内核诱导距离中的简单两项代替传统的欧氏距离, 将原有的空间像素映射到高维特征空间, 提高线性可分概率和计算速度; 利用图像像素之间的空间相关性, 通过引入改进的马尔可夫随机场对FCM目标函数进行修正, 提高算法的抗噪能力以及分割精度; 采用秃鹰搜索(BES)算法确定FCM的初始聚类中心, 提高算法的收敛速度, 同时避免算法陷入局部极值的情况。为验证改进FCM算法的性能, 选取划分熵、划分系数、Xie_Beni系数以及迭代次数作为评价指标, 并与近年来先进的图像分割算法进行对比。实验结果表明, 改进FCM算法具有更好的抗噪能力, 能得到更好的缺陷分割效果, 对工业生产中的冲压件缺陷检测有一定的应用价值。

关键词: 模糊C均值聚类, 工业应用, 冲压件缺陷, 内核诱导距离, 马尔可夫随机场, 秃鹰搜索算法