为了提高海上红外弱小目标检测的检测精度和实时性,提出了一种基于加权场景先验的红外弱小目标检测方法.该方法首先利用目标的稀疏特性以及海面场景的非局部自相关特性,将目标和背景的分离问题转化为恢复低秩和稀疏矩阵的鲁棒主成分分析(Robust Principal Component Analysis,RPCA)问题.之后,将海面背景的先验特征信息通过加权核范数的方式加入模型,加快算法中目标和背景图像块矩阵的分解速度.最后,通过引入交替方向乘子法(ADMM)算法进一步加速求解的迭代速度.实验结果表明: 该算法能有效地提高目标检测准确率,算法实时性较原算法提高了 120%.

To further to improve the detection accuracy and real-time performance of infrared small tar-get detection at sea,a new method based on weighted scene priors is introduced. Firstly,using the sparse characteristics of the target and the non-local self-correlation characteristics of the sea back-ground,the target-background separation problem is modeled as a robust low-rank matrix recovery problem. Moreover,the prior information on sea backgroundis addedinto the modelby weighted nu-clear norm to accelerate the decomposition of target and background images’matrix in the algorithm. Finally,the alternatingdirection method of multipliers (ADMM) is introduced to further to accelerate the iteration speed of the solution. The experimental results show that the proposed algorithm can effectively improve the accuracy of target detection. The real-time performance of the algorithm is improvedby 120% compared with the original algorithm.