基于视觉技术的煤岩特征分析与识别

Characteristic Analysis and Recognition of Coal and Rock Based on Visual Technology

为研究煤岩自动识别技术,提出了一种基于视觉技术的煤岩特征分析与识别方法。首先根据煤和岩石图像分析煤岩纹理差异;然后根据灰度共生矩阵分别计算煤和岩石纹理特征向量;最后选择计算出的纹理特征向量作为神经网络输入来分别识别煤和岩石2种情况。实验结果表明,煤和岩石纹理特征值差别较大,采用能量、对比度、相关性和熵作为特征向量均可实现煤和岩石自动识别,且以熵值作为特征向量的煤岩识别效果最好。

In order to research coal and rock automatic recognition technology, a method of characteristic analysis and recognition of coal and rock based on visual technology is proposed. Firstly,according to the coal and rock image, the texture difference between coal and rock is analyzed.Secondly,four texture feature values are calculated based on spatial gray co-occurrence matrix method.Thirdly,the coal-rock is recognized using neural network trained by texture feature. The result shows that there is a great difference of texture feature between coal and rock. Coal and rock can be recognized by using energy, contrast, correlation, entropy as feature vector. In addition, entropy has a better effect in recognition.