论文部分内容阅读
作为一种自然景物的模型,分形几何越来越多地受到了人们的关注。分形纹理分析是分形在图像处理中的应用研究的一个主要内容。注意到仅用分形维数特征不足以描述和区分不同的图像纹理,本文采用不同尺度下的方向性分维的变化曲线作为纹理图像的特征,对不同纹理的图像进行识别。实验表明这一方法是有效的。最后,本文讨论了该方法在雷达目标情况识别技术中的应用。
As a natural scene model, fractal geometry is more and more people’s attention. Fractal texture analysis is one of the main contents of the application of fractal in image processing. It is noticed that the fractal dimension feature is not enough to describe and distinguish different image textures. In this paper, we use the change curves of directional fractal dimension at different scales as the features of the texture image to recognize the images of different textures. Experiments show that this method is effective. Finally, this paper discusses the application of this method in radar target recognition.