基于DFBIR场的图像边缘提取的一种新方法

A New Method for Edges Detection Baseed on DFBIR Field Model

自然纹理的灰度图象是满足分数布朗随机场模型的．本文采用改进了的分形图象模型，即离散分数布朗增量随机场[2]（discretefractionalBrownianincrementalrandomfield，简记为DFBIR场）模型，通过求取图象局部分形特性参数H值，来提取图象的边缘。并且提出了一种有效确定分形特征参数H值二值化门限的新方法，采用该方法可以获取好的图象边缘．

The grey image of natural textures satisfy the model of fmetional Brownian random field. In this pa-per, an inproed fractal image model, which is discrete fractional Brownian incremental random field, isused andthe edges are detected by estimating the fractal parameter H. A new method for the determination of binary im-age threthod of parameter H is proposed and good edges can be obtained by using this method.