基于AP聚类算法的地板块纹理识别研究  

Research on Texture Recognition of Floor Based on AP Clustering Algorithm

作　　者：苑毓航 赵智超[1] YUAN Yuhang;ZHAO Zhichao(College of Information Science and Electronic Technique, Jiamusi University 154007, China)

机构地区：[1]佳木斯大学信息电子技术学院,佳木斯154007

出　　处：《微处理机》2018年第6期44-46,共3页Microprocessors

基　　金：佳木斯大学教育教学研究项目(2018JYXB-044)

摘　　要：为解决地板块纹理识别难度大,以及需要人工标注样本的问题,提出一种基于近邻传播聚类算法的地板块纹理识别方法。在详细介绍AP聚类算法的基础上,通过灰度共生矩阵获取地板块的纹理特征,包括对比度、能量、相关度、熵;将获取的图像特征输入AP聚类算法,更新吸引度与归属度函数来确定聚类中心数,实现地板块纹理的自动聚类和识别。实验结果表明,该方法辨识准确率高,不需要人为设置算法参数,优于传统的K-means聚类算法和K-medians算法,为地板块纹理识别研究提供了一个新的视角,具有一定的理论和实用价值。In order to solve the problem of difficulty in recognizing floor texture and the need to label samples manually,a floor texture recognition method based on neighbor propagation clustering algorithm is proposed.Based on the detailed introduction of AP clustering algorithm,the texture features of the floor slab,including contrast,energy,correlation and entropy,are obtained through gray level co-occurrence matrix.The obtained image features are input into the AP clustering algorithm,and the attraction and attribution functions are updated to determine the number of clustering centers and realize automatic clustering and recognition of floor texture.The experimental results show that this method has high recognition accuracy,does not need to manually set algorithm parameters,is superior to the traditional K-means clustering algorithm and K-medians algorithm,provides a new perspective for the study of floor texture recognition,and has certain theoretical and practical value.

关 键 词：纹理识别 AP聚类算法 灰度共生矩阵 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]