摘要
SAR图像目标检测方法没有固定的范式,是图像分析、理解和解译的基础。针对含有目标的SAR图像,提出一种基于Top-Hat变换和四叉树分解理论的SAR图像目标检测方法。对含有目标的灰度SAR图像,利用Top-Hat变换减小图像的突变,并使用图像的特征均值和特征方差值作为阈值进行动态阈值的四叉树分解。最后根据目标固有的特征对候选的若干区域选择一个真实的目标,并采用X波段SAR图像验证该方法能够有效地实现SAR图像目标区域的检测。
SAR image target detection method is a basis of SAR image analysis, understanding and interpretation and has no fixed mode. A method for SAR image target detection is proposed based on Top-Hat transform and quadtree decomposition theory for the SAR image with targets. For gray-scale SAR image with targets, Top-Hat transform is used to reduce the sahation of SAR image and the mean and the variance of image features are defined as dynamic thresholds for dividing SAR image using quadtree theory. Finally, targets of region-of-interests （ROIs） are selected using targets inherent features. And the method is veri- fied via the X-band SAR image and the result show that the method can detect the target area effectively.
出处
《现代电子技术》
2013年第3期30-33,共4页
Modern Electronics Technique
