摘要
为了解决声纳图像分类较困难的问题,提出一种基于分形理论的声纳图像识别方法。针对传统的分形维数作为图像纹理特征表示时缺少图像分布特征的不足,引入了多重分形的概念和广义维数,并对原图像和灰度差分图像重新提取了图像的纹理特征。通过对声纳图像的分类实验,结果证实该方法行之有效。
A method of sonar image recognition based on fractal theory is presented. The aim of this method is to solve the difficulty problem of sonar image recognition. Traditional fractal dimension methods lack image distribution information when they are used as image texture characters. Multi-fractal generalized dimension is put forward. Texture features of original images and gray differential images are extracted by using this method. In the end, the experiment of recognizing sonar images has been done and the result is satisfactory, which verifies the effect of this method.
出处
《微机发展》
2004年第4期55-57,共3页
Microcomputer Development
基金
十五"预研项目
