基于改进的奇异值分解和形态滤波的弱小目标背景抑制

摘要

红外弱小目标的复杂背景抑制一直是弱小目标检测与跟踪的一个难点。提出一种改进的奇异值分解和形态滤波Tophat变换相结合的红外弱小目标背景抑制算法。首先通过奇异值分解得到原红外图像的奇异值矩阵和左右奇异矩阵，然后通过对奇异值进行对数非线性变换，利用优化后的奇异值矩阵进行重构得到增强对比度的红外图像，最后利用形态滤波中的Tophat变换进行滤波达到背景抑制的目的。实验结果表明，该算法能够很好地实现红外弱小目标图像的背景抑制，并能使目标信号得到保存和增强。%Since the complicated background suppression is a difficulty for infrared dim small target detection and tracking, the background suppression algorithm for infrared dim small target based on the combination of improved singular value decompo⁃sition(SVD)and Tophat conversion in morphological filtering is presented. The singular value matrix and left and right singular matrixes of the original image are decomposed by singular value,and then the logarithm nonlinear transformation for singular val⁃ue is conducted. The optimized singular value matrix is reconstructed to obtain the infrared image of enhancing contrast ratio, and then the filtering for the image is conducted by Tophat conversion in morphological filtering to achieve background suppres⁃sion. The experimental results show that the algorithm can realize the perfect background suppression for the infrared dim small tar⁃get image,and preserve and enhance the target signal.