联合NSCT和SVD方法的红外弱小目标复杂背景抑制

NSCT Combined with SVD for Infrared Dim Target Complex Background Suppression

复杂背景抑制是天基红外预警系统中红外弱小目标探测技术的一个关键环节。为降低复杂背景下杂波干扰,提高目标检测精度,利用非下采样轮廓波变换(NSCT,non-subsampled contourlet transform)的多尺度分解及多方向分解特性以及图像矩阵奇异值分解(SVD,singular value decomposition)不同奇异值代表图像不同能量信息的特点,提出了联合NSCT和SVD的红外图像背景的抑制方法。首先依据非下采样轮廓波变换思想对红外原始图像进行多尺度多方向分解,得到与原始图像同样大小的不同尺度和不同方向上的子带图像,然后,利用奇异值分解的中序部分奇异值调整各子带图像矩阵系数以区分目标和背景杂波,最后对调整后各子带系数组成的矩阵施加NSCT逆变换,最终获得抑制背景处理后的图像。对比实验表明,该方法能够在低信噪比环境下有效抑制复杂背景及边缘,突显目标,降低虚警率。

Complicated background suppression is a difficult problem for the technique of the detection of the dim and small target by the SBIRS. To reduce the clutter interference and enhance detection precision, a new method based on NSCT combined with SVD is proposed for infrared complex background suppression, which utilizes the decomposed property about the multi-scale and directional details of NSCT and the property of the different singular value representing energy information of SVD in image array. Firstly, NSCT is adapted to decompose the input infrared image based on the multi-scale and directional details, which extracts multi-scale and directional sub-band images which equal to the original images. Then, the middle singular values based on the SVD are introduced to adjust the sub-band image array coefficients so as to differentiate target and background clutter signal. Finally, the suppressed image is reconstructed by the inverse NSCT. Contrastively experimental results show that the proposed algorithm has better performance in suppressing the complex background and its edges, making target protuberant and reducing false-alarm probability.