论文部分内容阅读
在传统的马尔可夫随机场(MRF)的图像建模方法基础上利用合成孔径雷达(SAR)图像的固有特性对Gibbs-MRF模型进行改进复原SAR图像,并进一步提出用数字形态学中连通性理论进行图像分割。在SAR图像像素空间的邻域内,估计最大后验概率(MAP)时引用Gamma分布代替传统的瑞利分布恢复数据,同时利用像素强度值相关性的连通模型将目标较好地提取出来。充分利用了SAR图像的数字形态信息和像素强度之间的相关性,得到了更好的分割效果。仿真实验说明本文方法是有效的。
Based on the traditional MRF (Markov Random Field) image modeling method, the intrinsic characteristics of Synthetic Aperture Radar (SAR) images are used to improve the SAR image of Gibbs-MRF model. Furthermore, it is further proposed to use digital morphological connectivity Theory of image segmentation. In the neighborhood of SAR image pixel space, the maximum a posteriori probability (MAP) is estimated to refer to the Gamma distribution instead of the traditional Rayleigh distribution to recover the data, and the target is well extracted by using the connectivity model of pixel intensity value correlation. Taking full advantage of the correlation between the digital morphological information and pixel intensity of SAR images, a better segmentation result is obtained. Simulation results show that this method is effective.