CAJ | 학술논문

　　针对SAR和光学两种异构传感器图像配准精度的局限性，提出一种新的分割方法及融合策略，最大限度地降低对配准精度的要求，提高了非精确配准下SAR和光学图像融合检测概率。首先对SAR和光学图像中的目标进行边缘和区域信息的提取，并以Ratio算子值为依据进行边缘连接，将图像分为有限个连通区域；然后通过模糊集判断进行两图像相近区域融合，以域内Ratio均值作为隶属度权值进行融合检测，在保证检测概率的前提下降低了虚警，获得了比单一传感器更精确、更完整的目标信息。
　　침대SAR화광학량충이구전감기도상배준정도적국한성，제출일충신적분할방법급융합책략，최대한도지강저대배준정도적요구，제고료비정학배준하SAR화광학도상융합검측개솔。수선대SAR화광학도상중적목표진행변연화구역신식적제취，병이Ratio산자치위의거진행변연련접，장도상분위유한개련통구역；연후통과모호집판단진행량도상상근구역융합，이역내Ratio균치작위대속도권치진행융합검측，재보증검측개솔적전제하강저료허경，획득료비단일전감기경정학、경완정적목표신식。
To overcome the limitation in image registration accuracy of SAR and optical sensors,a new segmentation and fusion method was put forward,which could enable the multi-sensor detection based on image fusion under the situation of low accuracy image registration and improve the detecting rate .Firstly,the edge/region information was extracted for the target in SAR and optical images .Then the edges were linked according to Ratio value,and the image was divided into several connected regions .Finally,the nearby regions in two images were fused,taking the mean value of Ratio value in regions as the weight factors of the fussy grade of membership .This algorithm can depress the false alarm rate while keeping the detection rate,and obtain more accurate and completed information of target than single sensor does .