针对合成孔径雷达(SAR)图像提出了一种带边缘惩罚和自适应权马尔科夫随机场(MRF)模型的快速分割算法。在MRF分割模型的能量函数中引入了边缘惩罚和自适应加权参数。边缘惩罚的引入能够减少边缘模糊从而更加精确地定位边缘。自适应权参数能够根据图像分割时收敛的阶段以及图像的局部场景自适应地调整能量函数中数据模型因子的权重,这有利于在均质区域改进分割区域的一致性,而在非均质区域则可保持图像的边缘和重要细节。针对所提出的能量函数提出了一种快速的非均质点跟踪优化算法。对合成和真实的SAR图像的分割结果表明,所提出的分割方法能提高分割的精度并显著减少运行时间。 Aiming at synthetic aperture radar (SAR) images, a fast segmentation algorithm with edge penalty and adaptive weighted Markov random field (MRF) model is proposed. Edge penalty and adaptive weighting parameters are introduced in the energy function of MRF segmentation model. The introduction of edge penalties can reduce edge blurring and locate edges more accurately. Adaptive weight parameters can adaptively adjust the weights of data model factors in the energy function according to the stage of convergence in the image segmentation and the local scene of the image, which is beneficial to improve the consistency of the segmentation region in the homogeneous region, while in the non-homogeneous region You can keep the edges and important details of the image. Aiming at the proposed energy function, a fast non-homogeneous point tracking optimization algorithm is proposed. The results of the segmentation of synthetic and real SAR images show that the proposed segmentation method can improve the segmentation accuracy and reduce the running time significantly.