基于鲁棒估计的微弱目标检测噪声抑制技术

抑制目标图像中的噪声,缩短目标的首次捕获时间(FirstCaptureTime,FCT),对于微弱目标检测和跟踪具有重要的意义。基于以下判断:a.任何检测系统总是工作在一定的信噪比(SNR)水平下;b.运动目标成像面积总是大于一个像素,分别设计了两个噪声抑制算子Sa(.)及Sb(.)。在噪声抑制过程中,首先使用鲁棒回归估计方法,估计出噪声模型参数,然后分别应用噪声抑制算子Sa(.)及Sb(.)对图像噪声进行抑制。试验结果表明,该方法可以有效抑制目标图像中的噪声,缩短目标的首次捕获时间,为下一步目标的检测和跟踪提供了有利条件。

Method of noise suppression based on robust regressive estimation for dim moving small target detection

A new method to suppress the noise in scene image and shorten the first capture time for dim small target detection and tracking is proposed. Two noise suppression operators, S_a() and S_b(), are introduced based on these assumption: a. all target detection and tracking system always works under some SNR condition; b. the projection pixels of the target is more than one pixel. The noise model parameters firstly by using robust regressive estimation are estimated, then the noise in target image by using S_a() and S_b() is suppressed. The simulation results show that the method can suppress the noise and shorten the first capture time of the target effectively, which turns the target detection and tracking to advantage.