基于UDWT与Snake模型的多时相SAR图像变化检测方法

MULTITEMPORAL SAR IMAGES CHANGE DETECTION METHOD BASED ON UDWT AND SNAKE MODELS

提出一种基于非下采样离散小波(UDWT)和Laplace分布假设下Snake模型的多时相SAR图像变化检测方法。该方法首先计算输入SAR图像的对数-比值图像。然后利用非下采样小波分解该对数-比值图像,得到其多尺度表达。最后迭代执行以下两步直到收敛:(1)利用最大似然法估计Laplace分布参数;(2)根据图像数据和上一步估计的参数演化当前曲线。实验结果验证了该方法的有效性。

A multitemporal SAR image change detection method based on undecimated discrete wavelet transform（ UDWT） and the Snake models under the assumption of Laplace distribution is proposed in this paper. In the method the log-ratio image inputting to SAR image is firstly calculated. Then the log-ratio image is decomposed by UDWT to obtain the multi-scale representation. The following two steps are iterated afterwards until the convergence is reached：（ 1） Estimating the parameters of the Laplace distribution using maximum likelihood method;（ 2） Evolving current curve according to the image data and the parameters estimated in last step. Experimental results verify the effectiveness of the proposed method.