标题:
基于岩屑图像的火山岩岩性识别Lithologic Identification of Volcanic Based on Cutting Figures
作者:
罗韬, 余艳梅, 滕奇志
关键字:
岩屑图像, 特征融合, 极限学习机, 火山岩识别Cutting Figures, Feature Fusion, Extreme Learning Machine, Volcanic Recognition
期刊名称:
《Journal of Image and Signal Processing》, Vol.5 No.2, 2016-04-08
摘要:
本文提出了一种基于岩屑图像的火山岩岩性识别方法,首先通过实验分析并确定了对火山岩岩屑识别率较高的颜色与纹理特征;其次通过对这些特征的组合研究,提出了基于线性加权的特征融合法;最后采用极限学习机作为分类器,对特征融合后的火山岩岩屑进行测试。最终实验表明,融合特征对于火山岩岩屑能达到92.05%的识别率,为火山岩的岩性识别提供了一种可靠的参考依据。
In this paper, a new method based on features fusion for volcanic rocks lithology recognition is proposed. First, the color and texture with higher recognition rate are analyzed and determined through experiments. Secondly, a method based on feature fusion is proposed by combining the features into a new integration feature. And finally the extreme learning machine is used as the classifier. Experiments show that the recognition accuracy of integration features is up to 92.05%. This method provides a reliable reference for lithology identification of volcanic.