分形方法在SAR图像分割中的应用

SAR图像中人造地物和自然地貌很容易在某些特征上出现类似,使得利用传统方法提取的特征进行分割往往不可靠.相比自然地貌,人造地物显得比较规则,SAR图像的这一特点使得基于分形特征的分割算法成为可能.基于数学形态学理论提出了一个变结构元素的形态学分形维数估计方法,对复杂SAR图像进行分割.实验结果表明,该方法计算量小,可充分利用SAR图像的纹理信息.

SAR image edge detection based on morphological estimation of fractal dimension

Since certain features of the artificial ground objects are similar with that of natural physiognomies in SAR image, traditional segmentation algorithms are often not reliable for SAR image contains abundant information. Artificial ground objects look more regular, while natural physiognomies represent certain scrambling. Thus it is possible to use fractal feature for image segmentation of SAR image. An estimation method of fractal dimension is put forward using a changeable structural element based on morphological principal to detect the edge of complicated SAR image. The experiment shows that this method can make full use of texture information of SAR image and has small computation expense.