基于局部窗口的红外弱小目标检测方法

Infrared Dim and Small Objects Detection Based on Local Window

在分析红外图像中弱小目标检测方法的基础上,提出了一种基于局部窗口检测红外弱小目标的新方法。该方法利用图像的局部信息预测背景,直接根据当前像素点与背景像素点的灰度差及固定的分隔阀值检测目标,可以在低信噪比、强度变化剧烈的图像中获得较高的检测率与较低的虚警率。实验结果表明了该算法的有效性。

With the analysis of dim and small objects detection in infrared image,a new method based on local window was presented.The method is： use pixels in local window to predict background,and detect object pixels by const threshold.Experimental results show the validity of the method and demonstrate that the method can achieve higher detection probability and lower false alarm rate in low SNR and strong background image.