基于MATLAB的煤岩图像识别算法的研究

煤岩成分识别的目的是识别粘结剂、镜质组和非镜质组。为了能够快速准确地实现煤岩成分识别，采用受限的自动阈值法分割粘结剂，并根据图像每个小窗口中粘结剂成分的比例去除噪声。随后重点介绍了均匀度法识别镜质组，并对这种方法作了理论上的研究和实验上的性能比较。结果表明，均匀度法提取的镜质组识别率和图像信噪比均有了很大的提高，运算速度也大大加快，是综合效果较好的一种方法。

Research of Coal Petrography Image Recognition Algorithm Based on MATLAB

The aim of coal petrography composition recognition is to identify the padding, vitrinite and non- vitrinite. To be able to identify the composition quickly and accurately, a limited auto threshold method is adopted in this paper to identify the padding and eliminate the noise by the ratio of padding in each little window of the image. The PCA and uniformity quality algorithm are used to i- dentify vitrinite in this article, then the theoretical research and practical comparison on these two methods are made. The results show that the identify rate and image SNR of vitrinite that picked up by using the uniformity quality algorithm are greatly improved than that by using the PCA , and the computing speed is also much faster. So the uniformity quality algorithm is a better method with better effect.