基于模糊C均值聚类的纺织品印花图像分割

Textile print image segmentation based on FCM

针对纺织品印花精度检测的关键技术问题,主要围绕纺织品印花图像的有效分割展开研究工作,提出了基于距离修正和隶属度修正的FCM算法。为了确保FCM算法的有效性和精确性,对其参数的选取和算法结构进行了深入的研究,结合参数的自适应初始化方法和FCM算法的修正方法,给出该算法对纺织品印花图像的分割效果图。结果表明,该算法对纺织品印花图像有良好的分割效果,并且对由光和织物纹理引起的干扰有一定的抑制作用。

Focusing on the key issue in textile print precision measurement,the paper studies an effective method for image segmentation of textile print pattern.FCM algorithm is proposed,which is based on distance correcting and subordinative degree correcting.In order to ensure the effectiveness and accuracy of FCM algorithm,deep study of the parameter selection and algorithm structure is conducted.The self-adaptive parameters initialization method and FCM algorithm correction method are investigated by experiments.Finally,the results of image segmentation of textile print pattern by using this algorithm are demonstrated.The experimental results show that the algorithm has good effect on image segmentation of textile print pattern,and has certain inhibition effect on the interference caused by light and fabric texture.