云天背景下的红外弱小目标检测算法

Detection algorithm of weak infrared point targets under cloud background

根据点目标在红外系统中的成像特点,提出了一种红外弱小目标快速检测算法,首先进行下采样来减小图像大小;再运用"梯度增大法"增强图像对比度;然后与背景预测图像做差值运算,用最佳门限从差图中得到候选目标集;最终利用相临帧的相关性在后续帧处理中确定目标。仿真试验表明本文算法能从复杂云层背景中快速有效地检测出目标。

According to the imaging characteristic of the point target in infrared system, a fast detection algorithm to weak infrared targets is proposed. Firstly, the down sampling method is applied to reduce the image size. Secondly, ＂increasing gradient method＂ is used to enhance the image contrast. Then subtraction operation is performed between the disposed image and background forecasting image and a muster of candidate targets are selected form the subtraction images according to the best threshold. Finally, real target is obtained from the subsequent frame processing based on the correlation between near frames. The simulation shows that the algorithm is capable of detecting real target form the complex cloudy layer background quickly and efficiently.