论文部分内容阅读
针对合成孔径雷达(SAR)图像特有的乘性噪声和非恒虚警统计特性很难正确提取目标边缘的问题,提出了在指数加权均值比(ROEWA)算子基础上寻找自适应的最佳局域Gabor滤波器进行目标边缘提取的方法。利用Gabor滤波器具有的多方向特性确定边缘方向,然后用最大似然估计纠正错误边缘方向,重新结合视觉细胞倍频程计算出Gabor函数的最佳局域滤波参数,提取出SAR图像的正确边缘。实验表明,该方法取得很好边缘提取效果,并且后期分割出的目标更符合实际目标形态,具有较强的通用性。
Aiming at the problem that it is difficult to correctly extract the edge of the target by multiplicative noise and non-constant false alarm statistic characteristic of synthetic aperture radar (SAR) images, an adaptive best-fit game based on the exponential weighted mean ratio (ROEWA) operator is proposed Domain Gabor Filter Method for Target Edge Extraction. The direction of the edge is determined by using the multi-directional characteristic of the Gabor filter, and the direction of the false edge is corrected by maximum likelihood estimation. The optimal local filtering parameters of the Gabor function are calculated by reassembling the octave of the visual cell, and the correct edge of the SAR image is extracted . Experiments show that this method has a good edge extraction effect, and the target segmented later is more in line with the actual target morphology, with a strong versatility.