基于图像处理的煤矸识别方法  

Coal and Gangue Identification Method Based on Image Processing

作　　者：田冬艳 丁苏凡 郭星歌 TIAN Dongyan;DING Sufan;GUO Xingge(School of Information and Control Engineering,China University of Mining and Technology,Xuzhou 221116,China)

机构地区：[1]中国矿业大学信息与控制工程学院,江苏徐州221116

出　　处：《煤炭技术》2022年第3期201-204,共4页Coal Technology

基　　金：国家重点研发计划资助项目(2018YFC0808302)。

摘　　要：综合考虑煤和矸石图像的灰度特征和纹理特征,在对图像进行滤波、增强等预处理后,筛选出灰度直方图的均值、峰值,GLCM的能量、对比度和熵,Tamura纹理的对比度这6个特征组成特征向量,送入LS-SVM进行识别。研究结果表明:基于3种特征结合的LS-SVM煤矸识别有效地提高了识别率。The gray feature and texture feature of coal and gangue images are considered comprehensively,after filtering and enhancing images, six features, including the mean value and peak value of gray histogram, the energy, contrast, entropy of GLCM, and the contrast of Tamura texture, are selected to form feature vectors, which are then sent into LS-SVM for recognition. The results show that, LS-SVM based on the combination of three features can effectively improve the recognition rate of coal and gangue.

关 键 词：煤矸识别 图像处理 灰度特征 纹理特征 最小二乘支持向量机 

分 类 号：TD94[矿业工程—选矿]