矿石的显微图像成分分析是一项具有挑战性的应用研究课题，针对所要求的对某种矿石进行自动识别的任务，提出了一种使用改进的统计假说方法对网孔状纹理成分进行识别的方法；并提出了一种采用Sobel变换、骨架提取以及Hough变换相结合的技术进行针板状纹理成分识别的方法，实验结果表明，该方法是有效的。

The ingredient analysis of ore microscope image is a challenging research area. To accomplish automatic recognition task on a kind of ore, this paper proposed an improved statistical hypothesis method for mesh texture ingredient recognition; and with Sobel edge detection, Skeleton Extraction, Hough Transformation another method is proposed on the recognition for needle like or striation texture ingredient. The experimental results show that the methods proposed in this paper are efficient.