【作者】 王一帆； 周丹； 屈玉福；

【Author】 WANG Yi-fan;ZHOU Dan;QU Yu-fu;School of Instrumentation Science and Opto-electronics Engineering,Beihang University;

【机构】 北京航空航天大学仪器科学与光电工程学院；

【摘要】 在多模式目标探测技术中,利用可见光图像和红外图像的匹配融合是解决单一传感器不能全天候、全天时工作和图像信息少的有效手段。本文根据红外图像与可见光图像的特点,首先在图像预处理阶段设计了基于大津阈值的红外图像分割算法,得到完整的目标红外图像;然后设计并计算可见光图像和红外图像间的仿射变换矩阵,并采用逆向计算的方法避免配准图像出现的空洞和网格的现象。最后经实验验证,本文所设计的可见光和红外图像匹配融合技术有很高的匹配精度。更多还原

【Abstract】 Single sensor cannot accomplish an all-weather work or an all time work,and it provides few image information.So the registration and fusion of visible image and infrared image in multi-modal object detection techniques is an efficient measure to solve these problems.In this paper,according to the characteristics of visible image and infrared image,a segmentation algorithm of infrared image based on OTSU method is firstly designed to get the complete target Then the affine transformation array between visible image and infrared image is designed and calculated,which avoids the cavity and grid in registration by taking the inverse calculation.Finally,the experiment shows that the technique on the registration and fusion of visible image and infrared image in this paper has an accurate registration result.更多还原