CAJ | 학술논문

针对单帧红外弱小目标检测中背景抑制残差图中的目标检测问题，提出一种基于窄波段像素色比的残差图融合方法。首先在传统背景抑制获得残差图的基础上，通过计算窄波段图像的像素色比，选取适当的恒定分割率实现目标和杂波的分离，依照特定的融合准则获得信噪比较高的融合残差图；然后应用基于目标体积检测的方法对融合图像中目标能量进行集中，以获取信噪比更高的结果图。该方法的优点在于，弥补了背景抑制后的图像直接阈值分割时受杂波影响大的缺陷，能够有效降低虚警率。
침대단정홍외약소목표검측중배경억제잔차도중적목표검측문제，제출일충기우착파단상소색비적잔차도융합방법。수선재전통배경억제획득잔차도적기출상，통과계산착파단도상적상소색비，선취괄당적항정분할솔실현목표화잡파적분리，의조특정적융합준칙획득신조비교고적융합잔차도；연후응용기우목표체적검측적방법대융합도상중목표능량진행집중，이획취신조비경고적결과도。해방법적우점재우，미보료배경억제후적도상직접역치분할시수잡파영향대적결함，능구유효강저허경솔。
The aim of this paper is to solve the target detection after the background suppression in infrared small target detection.A method based on the color ratio of different narrow wavebands is proposed to fuse the residual images.There are four main procedures in the whole detection:first,the background for residual images is suppressed;second,the fused residual images with improved SNR are obtained by color ratio calculation;third,the concentrated target energy in the images is obtained by employing the target volume detection method bringing a higher SNR of the image;finally,threshold segmentation would find the targets more precisely compared with traditional segmentation.The proposed method has advantages of being unacted on the clutter and reducing the false alarm rate.