论文部分内容阅读
结构化背景抑制是红外弱小目标检测技术的一个难题。根据红外图像中目标和背景信号的特性,在定义Gabor函数为点扩展函数的基础上,提出了一种基于全变差滤波的背景抑制算法,该算法将图像的背景抑制转化为求解全变差滤波模型最小化,以便最终实现对红外图像的背景抑制。对真实的红外图像序列进行实验,并与小波域滤波算法进行了比较。几组实验结果表明,对在结构化背景下的红外弱小目标背景来说,从主观视觉和数值指标来看,该算法具有良好的抑制效果,且运算量较小,便于实时实现。
Structured background suppression is a difficult problem for infrared weak target detection. Based on the characteristics of target and background signal in infrared image, a Gabor function is defined as a point spread function, and a background suppression algorithm based on total variation filtering is proposed. The algorithm suppresses the background suppression of the image into the solution of total variation The filtering model is minimized to finally achieve background suppression of the infrared image. Experiments on real infrared image sequences are carried out and compared with the wavelet domain filtering algorithm. Several groups of experimental results show that this algorithm has good suppression effect on subjective visual and numerical indexes for the weak infrared target background in the structured context, and the computation is small, which is convenient for real-time implementation.