图像颜色特征提取在铁谱图像分类及磨粒识别中的应用研究  

Application Study of Color Feature Extraction on Ferrographic Image Classifying and Particle Recognition

作　　者：陈桂明[1] 谢友柏[1] 江良洲[2] 

机构地区：[1]西安交通大学,西安710049 [2]第二炮兵工程学院,西安710025

出　　处：《中国机械工程》2006年第15期1576-1580,共5页China Mechanical Engineering

基　　金：图像信息处理与智能控制教育部实验室开放基金资助项目(TKLJ0307)

摘　　要：铁谱图像所包含的彩色信息对磨粒识别、磨损形式分析非常重要。对图像分类技术进行研究,主要对铁谱图像的颜色特征进行研究,提出了聚类树分析、模糊聚类技术与统计分析相结合的定量研究方法,对铁谱图像进行背景、磨粒区域分割,以获得可进行定量分析的磨粒。计算的颜色特征为铁谱图像的进一步处理和识别以及磨粒的机器自动识别、磨损形式分析奠定了基础。Color information of ferrographic image is important to particle identification and wear condition analysis. On the basis of studying on image classifying, this paper mainly studied color feature extraction of ferrographic image. The clustering tree, fuzzy clustering technique and statistic analysis were used in quantitative segmentation of background and wear particles of color ferrographic image. Calculated color features of background and different wear particles can be used for farther processing and recognizing of ferrographic image, and can also be used in particle machine automatic recognition and wear condition analysis.

关 键 词：铁谱技术 油样分析 图像分类 特征提取 模糊聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]