图像的纹理特征检测技术在互联网色情图像监测以及Web图像过滤等方面有着广泛的应用。通过阐述皮肤的纹理属性和纹理分析方法引出基于灰度共生矩阵的纹理检测技术，并且通过具体的实验给出了相关参数的取值，具体包括共生矩阵计算主要涉及的灰度级数和位移量，以及纹理特征度量值的选择。实验结果表明，反差和能量是区分皮肤区域和非皮肤区域的两个最佳纹理度量指标。

Image’s texture feature detection has many applications in the pornographic images monitoring and web images filtering. This paper described the methods of texture analysis and the texture characteristics of the skin, then gave the texture detection method based on gray level co-occurrence matrix. Through the experiments, the paper inferred the key parameters of the grey level number and the offset value in the grey level co-occurrence matrix computing. The experiment conclusion shows that contrast and energy are the best texture indexes which could distinguish the skin area and non-skin area.