基于光流估计和自适应背景抑制的弱小目标检测

针对复杂云背景下的弱小目标探测,提出了一种基于光流估计和自适应背景抑制相结合的弱小目标检测算法.首先根据红外图像中云的移动规律,对云背景下的红外图像进行光流分析,提取运动云区.在光流场的计算中结合了云运动的特点以及光流方程的两个约束条件,对传统的基于梯度的光流法予以改进.同时发现移动云区对目标探测的影响较大,为了抑制移动云区对弱小目标的干扰,提出了自适应抑制复杂背景的算法,在光流场分析提取的移动云区中,利用代表背景复杂程度的背景因子,自适应调整分割阈值,抑制复杂背景的干扰.这样只在容易引起虚警的移动云区进行背景抑制处理,简化了计算量,降低了云区对弱小目标的干扰,减少了虚警和误判.实验结果表明该算法可以显著减少云区造成的虚警,并且能够探测出弱小目标.

A Detection Algorithm for Dim and Small Infrared Target Based on the Optical Flow Estimation and the Adaptive Background Suppression

In connection with the infrared target detecting under complex cloud backgrounds,a small target detection algorithm based on the optical flow estimation and adaptive was put forward.Fistly,the infrared image under cloud backgrounds was analyzed based on the optical flow and the cloud movement was extracted.The traditional gradient-based optical flow was proved based on the characteristics of cloud movement and the two constraints of the optical flow equation in the calculation of optical flow.And,it was found that the cloud movement had a greater impact on the target detection,the algorithm of adaptive suppress the complex backgrounds was introduced to reduce the interference of the cloud.the background factors on behalf of background complexity was used to adjust the threshold in the cloud region extracted in the optical flow analysis.In this way,the background suppresion was produced in the cloud movements.It would easily cause false alarm and the algorithm simply the calculation,and reduce the influence of the cloud and the false alarm.Experimental results show that this algorithm can detect small targets and significantly reduce the false alarm caused by the cloud area.