强杂波中双波段目标检测新算法

对海空背景下的电视和3～5μm红外目标提出了一种采用多向梯度表决融合的检测方法：依据背景图像中点目标的奇异特性，得到图像的多向剃度检测结果，并进行表决融合。通过实测的电视和红外图像仿真，结果表明，采用“and”判决规则的多向梯度表决融合的检测算法消除了云层、海浪、海天线等背景干扰，在实现高检测概率的同时，可以达到较低的虚警概率。该算法特别适用于低帧率的红外警戒系统，可检测信杂比为1的点目标。

New Detection Algorithm of Weak Targets on Double Bands Under Strong Clutter

A new detection algorithm based on multi-orientation gradient voting fusion for the TV and 3 infrared target under the sea background were proposed. Firstly, the target is detected through the multi-orientation gradient of image based on the target singularity. Then the result of the detection of TV and infrared image is fused through the voting theory. The computer simulation results show mat the algorithm based on multi-orientation gradient voting fusion is available and gives better result. The algorithm is suited especial for surveillance system of low frame rate and can detect the point targets with signal to clutter ratio equal to 1.