SAR图像多尺度积增强的目标检测算法

合成孔径雷达（SAR）成像系统的热噪声和海杂波严重影响SAR图像自动目标检测的性能,去噪和均匀背景杂波是提高SAR图像目标检测性能的重要课题。根据SAR图像噪声功率一般存在于信号小尺度,没有跨尺度特征,而目标信号的边缘具有跨尺度的特点,本文提出了一种多尺度积信号增强和去噪的SAR图像船舰目标检测算法。本算法对SAR图像进行小波变换,应用多尺度积在小波域增强SAR图像船舰信号和均匀背景杂波,再对SAR图像进行目标检测。ERS SAR图像用于验证本文算法。仿真实验结果表明,新算法同传统的双参数CFAR检测算法、基于K-分布背景杂波的检测算法以及基于小波软阈值增强的检测算法相比,在虚警数和品质因数性能指标上均优于后几种检测算法。

SAR Image Enhancement Using Multiscale Products for Targets Detection

Since the heat noise of Synthetic Aperture Radar(SAR) imaging system and sea clutter severely affect the performance of automatic targets detection in SAR images,removing noise and homogenizing sea clutter in SAR images to improve performance of targets detection is a challenge for researchers.Considering the noise power exists small scale and the boundaries of target signal exist cross-scale in SAR images,we propose a novel scheme which enhances signal and removes noise based on multiscale products method to detect ship targets.The algorithm applies wavelet transform to SAR image and enhances ship signal and homogenizes background clutter of SAR images in wavelet domain using multiscale products.And then detects ship targets in enhanced SAR images.ERS SAR images are used to test our detection algorithm.The simulating results show that the new detector improves detection performance when it is to compare the two-parameter constant false alarm detector and the detection algorithm based on K-distribution.To estimate the efficacy of our enhancement algorithm,we compute computation complexity of the algorithm and a popular enhancement algorithm which is based on wavelet soft threshold.It shows that our algorithm is easier to implement using hardware.The two enhancement algorithms are applied to SAR image ship targets detection.The detected results show that the detection performance of our algorithm is better than the later.