阈值法是图像分割的一种重要方法, 在图像处理与识别中广为应用. 提出了一种基于灰度2梯度共生矩阵

模型和最大熵原理的自动阈值化方法. 该方法不仅利用了图像的灰度信息, 而且也利用了梯度信息, 通过计算基

于灰度2梯度共生矩阵的二维熵, 并使边缘区域的熵最大来选择阈值向量. 仿真结果显示, 该算法比其他二维熵方

法效果更佳.-Th resho lding is an impo rtant fo rm of image segmentat ion and is used in image p rocessing and recog2

nit ion fo rmany app licat ions. In th is paper, an automat ic app roach fo r th resho lding based on gray2level gradient co2

occurrence mat rix model and the maximum ent ropy p rincip les is p ropo sed. Th ismethod ut ilizes the info rmat ion of

bo th gray level and gradient in an image. In th is app roach, the th resho ld vecto r is selected th rough evaluat ing two2

dimensional ent rop ies based on the gray2level gradient co2occurrence mat rix and maxim izing the edge region en2

t rop ies. It is found that the p ropo sed app roach perfo rm s bet ter than o ther 2D ent ropy methods.