论文部分内容阅读
SAR图像极低的信噪比以及乘性噪声给SAR图像的边缘检测带来了较大的困难。提出了一种针对SAR图像边缘的自适应贝叶斯检测方法。该方法利用广义高斯马尔可夫随机场作为局部均值的先验概率分布模型,利用贝叶斯准则推导了局部均值的最大后验概率估计。广义高斯马尔可夫随机场模型参数估计和局部均值估计采用联合迭代技术进行求解。边缘检测器的参数采用接收机操作性能曲线和卡方检验进行选择。基于实测SAR数据的仿真实验结果表明,本文的边缘检测算子是有效的,并优于已有的SAR图像边缘检测算子。
The extremely low signal-to-noise ratio of SAR images and multiplicative noise bring more difficulties to the edge detection of SAR images. An adaptive Bayesian detection method for SAR image edge is proposed. The method uses the generalized Gaussian Markov random field as a priori probability distribution model of local mean and deduces the maximum posteriori probability estimation of local mean using Bayesian criterion. The generalized Gaussian Markov random field model parameter estimation and local mean estimation are solved by joint iteration technique. Edge detector parameters using the receiver operating performance curve and chi-square test to choose. The simulation results based on measured SAR data show that the edge detection operator in this paper is effective and superior to the existing SAR image edge detection operator.