基于Contourlet及目标特性分析的弱小红外目标检测

Small infrared target detection based on Contourlet transformation and target characteristic analysis

研究了一种复杂背景下红外小目标检测的新方法。根据远距离红外弱小目标在相平面上成像的各向同性特征以及目标与背景边缘分离困难的现状,提出了一种基于Contourlet及目标特性分析的弱小红外目标检测算法。首先采用Contourlet变换对图像中的低频起伏背景及边缘与随机噪声和目标进行分离;然后对红外目标进行基于各向同性特征的目标提取;最后对目标进行分割、检测以及定位,并输出目标信息。通过与一些传统方法进行基于主观视觉和客观指标的对比发现,该算法可有效分离目标与低频背景、随机噪声及背景边缘,对红外图像基于弱小目标检测背景处理时,效果较好。

A new method of small infrared target detection under complex background is studied in this paper.The imaging of long-range infrared dim and small targets on phase plane has isotropic characteristics.Separating target from background edge is very difficult.Then,a dim and small infrared target detection algorithm based on Contourlet and target characteristic analysis is proposed.First,the contourlet transform is used to separate the low-frequency fluctuating background and edge from the random noise and target.Then the infrared target is extracted based on isotropic feature.Finally,the target is segmented,detected and positioned,and the target information is output.By comparing with some traditional methods and methods proposed in literature based on subjective vision and objective indicators,the results show that the algorithm can effectively separate the target from low-frequency background,random noise and background edges,and has a better effect in infrared image background detection based on small and dim targets.