针对红外复杂背景下的弱小目标检测难题, 提出一种基于背景自适应的多特征融合的复杂背景下弱小目标的检测算法。首先, 通过对红外图像进行空域滤波去除孤立噪声点, 并利用恒虚警率分割消除大面积平稳背景, 获得疑似目标集。然后融合红外图像的背景信息、弱小运动目标的灰度特征、目标与周围像素的方向梯度特征等多个典型特征, 消除疑似目标集中的大部分假目标, 最后运用运动特征获取真实目标的轨迹, 最终实现复杂背景下的红外弱小目标的检测。实验表明: 该算法能实现复杂背景下低信噪比的红外弱小目标快速检测, 具有检测概率高, 算法速度快, 鲁棒性好的特点。

In order to solve the problem of infrared dim and small target detection under the complex background of infrared, a detection algorithm for dim and small targets in complex background based on background adaptive multi feature fusion was proposed. Firstly, the isolated noise points are removed by spatial filtering of infrared images, and the constant false alarm rate was used to eliminate large area stationary background and obtain a suspected target set. And then most of the false targets in the suspected target were eliminated by combining the background information of the infrared image, the gray feature of the dim and small moving target and the direction gradient feature of the target and the surrounding pixels. Finally, the detection of infrared dim and small targets in complex background was realized. The experiment shows that the algorithm can realize fast infrared dim and small target detection with low signal to noise ratio under complex background, which has the characteristics of high detection probability, fast speed and good robustness.