通过研究图像邻域相似关系，提出了“邻域相似比”的概念。邻域相似比能够表征图像邻域颜色的相似程度。然后以此为核心构建了图形图像分类模型，在Matlab环境中编译算法进行了仿真实验。结果表明，利用邻域相似比进行图形图像分类，算法简单快捷，运算成本低，正确率高。

Similarity relationships among neighborhood pixels of images were studied and a new concept "neighborhood similar ratio" was put forward. Neighborhood similarity ratio can represent similarity degree of neighborhood color of image. Then an image classification algorithm was proposed in terms of it and simulated in Matlab. The result showed that the image classification method based on neighborhood similarity ratio is fast and efficient with low cost.