低信噪比红外图像小目标的检测

文章提出了一种新的自适应平滑滤波方法,并据此设计自适应高斯核平滑滤波来抑制红外图像中的随机噪声。在此基础上依据小目标是孤立亮斑的特点设计出三阈值起始点提取方法,然后用有阈值限制的区域生长和区域合并的方法恢复没有被提取出来的面目标的内部和弱边缘。实验表明本方法的正确性和有效性。

Detection of Small Targets of Low SNR Infrared Images

The paper proposes a new method of adaptive smooth filter , based on that to design a Gaussian kernel adaptive smooth filter so as to suppress the random noise spread on the infrared image. Then a three threshold method is designed to extract the starting points according to the fact that the small objects in the infrared image are separated and have relatively high brightness. Consequently ,the other part of small objects containing dim edges and internal points are determined by the threshold limited region growth and merge method. Simulation results show that the methods demonstrated are effective and efficient .