一种基于PCA与RBF-SVM的煤岩显微组分镜质组分类方法

发布时间：2018-04-09 13:45

本文选题：煤　切入点：显微组分　出处：《煤炭学报》2017年04期

【摘要】：在分析煤岩镜质组显微组分特点的基础上,针对其结构复杂、特征量多且相互交织从而影响分类准确性等问题,提出一种基于主成分分析(PCA)的煤岩显微组分镜质组分类方法。首先根据镜质组显微图像中各组分呈现的条状、团块、颗粒等纹理特点和亮度差异,采用基于灰度共生矩阵的能量、熵、惯性矩、局部平稳性等纹理特征量和基于灰度分布统计的亮度比、均值、均方差、三阶矩偏度等亮度相关特征量对其进行描述,构成初始特征量集;再采用主成分分析法对初始特征量集进行进一步的抽取;最后构建基于径向基函数的支持向量机(RBF-SVM),采用积累贡献率较大的主成分作为分类参量实现镜质组的自动分类。实验结果表明:纹理和灰度统计特征可有效刻画煤岩镜质组显微组分;采用PCA对初始特征进行抽取之后,用于分类的特征空间维数大幅度降低,分类算法的泛化能力增强,分类的准确率显著提高。
[Abstract]:On the basis of analyzing the characteristics of vitrinite macerals of coal and rock, a classification method of vitrinite of coal and rock based on principal component analysis (PCA) is proposed, aiming at the problems of complex structure, large amount of characteristic and intertwined with each other, thus affecting the classification accuracy of coal and rock vitrinite.Firstly, according to the texture characteristics and luminance difference of each component in vitrinite microscopic image, the energy, entropy, moment of inertia based on gray level co-occurrence matrix are adopted.The local stationary texture feature and the luminance ratio, mean value, mean variance, third order moment deviation and other luminance correlation characteristics based on gray distribution statistics are described to form the initial feature set.Then the principal component analysis (PCA) is used to extract the initial feature set. Finally, the support vector machine (RBF-SVMN) based on radial basis function (RBF) is constructed, and the principal component with large cumulative contribution is used as the classification parameter to realize the automatic classification of vitrinite.The experimental results show that the texture and grayscale statistical features can effectively depict the macerals of coal and rock vitrinites, and after the initial features are extracted by PCA, the dimension of the feature space used for classification is greatly reduced, and the generalization ability of the classification algorithm is enhanced.The accuracy of classification is improved significantly.
【作者单位】：安徽工业大学电气与信息工程学院;安徽工业大学煤的洁净转化与综合利用安徽省重点实验室;安徽工业大学冶金减排与资源综合利用教育部重点实验室;
【基金】：国家自然科学基金资助项目(51574004) 安徽省自然科学基金资助项目(1208085ME67)
【分类号】：TQ533;TP391.41

【相似文献】