红外弱小目标检测与识别一体化方法研究

Researches on the Integration of Infrared Dim and Small Targets Detection and Recognition

针对天基红外预警卫星获得的红外图像的特点，分析了传统的红外弱小目标的检测与识别方法并指出了其存在的检测时间长，虚警概率高等不足之处。根据导弹预警系统对目标检测的实时性、快速性和准确性需求及导弹尾焰红外辐射光谱特性，提出了一种窄带高光谱的红外弱小目标检测与识别一体化方法。仿真实验采用了一个典型的红外背景图像，根据两类典型导弹尾焰在2．7～2．9μm波段范围内的红外辐射信息，生成了20幅SNR为2．0～2．5的红外图像，结果验证了该方法能快速准确的检测并识别导弹目标。

Aiming at the characteristic of the infrared images gained from the SBIRS, the paper analyzed the traditional method of infrared dim and small targets detection and recognition, then pointed out its detection time was long and the deficiency of high false alarm probability. An integration method of detection and recognition of narrowband multi-spectral was put forward based on the desire of the real time, quickness and accuracy of missile early warning system and the spectrum features of the missile plume infrared radiation. The simulation experiment used a typical infrared background image, according to the infrared radiation information of the two kinds of typical missile plume between 2.7-2.9 μm, generated 20 frames infrared images with SNR between 2.0-2.5. The result validated that such method can rapidly and accurately detect and identify the missile targets.