目前常用的无损图像压缩方法都是利用相邻数据组合除去数据中存在的相关性。既然局部或相邻像素间存在一定的灰色关联度 [1 ] ,而非线性灰色模型又可以从邻近像素中提取灰色关联 ,因此提出了一种快速灰色预测模型(FGMP) ,并构建非线性预测模式 ,可以在灰度图像无损压缩中实现预测误差编码。运用该算法进行了图像压缩 ,并得出了一些结论 The commonly used nondestructive image compression methods use the combination of adjacent data to remove the correlation existing in the data. Since there is a certain degree of gray relational between local or neighboring pixels [1], and non-linear gray model can extract gray relation from adjacent pixels, a fast gray prediction model (FGMP) is proposed and a nonlinear prediction Mode, predictive error coding can be implemented in gray-scale image lossless compression. The algorithm is used for image compression, and some conclusions are drawn