基于多光谱和全色影像的精分波段融合算法研究  

Research on Multi-spectral and Panchromatic Image Fusion Approach Using Precisely Resolved Band

作　　者：张艳秋 李艳芳 ZHANG Yanqiu;LI Yanfang(School of Surveying&Mapping and Geographic Information of Lanzhou Resources&Environment Voc-Tech University,Lanzhou 730000,China)

机构地区：[1]兰州资源环境职业技术大学测绘与地理信息学院,甘肃兰州730000

出　　处：《测绘与空间地理信息》2023年第3期73-75,82,共4页Geomatics & Spatial Information Technology

基　　金：2021年甘肃省高等学校创新基金项目--无人机点云数据深度学习分类方法在高压输电线路中的应用研究(2021A-252);兰州资源环境职业技术学院2020年院级科研能力提升项目--无人机多光谱传感器交叉定标研究(Y2020C-06)资助。

摘　　要：图像融合技术在很多领域得到了广泛应用,但是仍然不能满足人们的需求。本文基于多光谱图像和全色图像,提出了一种精分波段的融合算法。首先判断多光谱与全色图像像元之间的重叠方式,以灰度值跨度为依据判断像元类型是否相同,依据像元灰度差异将数据进行分组,通过目标方差判断提取差异像元位置,并可以设定阈值提高不同地物目标类型精度。该方法通过土地类型面积计算实验进行验证,提高了全色影像建筑物和农田目标的分类精度,验证了方法的有效性。Image fusion technology has been widely used in many fields,but it still has development space.In this paper,the fusion algorithm of a precisely resolved band has been proposed based on multi-spectral images and panchromatic images.The algorithm first determines the overlapping mode between the multi-spectral and the panchromatic image cells,and then the gradation value is based on whether it is the same as the determination cell type.The data is packetized according to the varying grayscale difference,and the extraction difference pixel position is judged by the target variance.The threshold can be set to improve the intensive type accuracy of different type objects.This method is verified by the land type area calculation experiment,and it can enhance the classification accuracy of the panchromatic image building and farmland objectives.This result verified the effectiveness of this method.

关 键 词：多光谱图像 全色图像 图像融合 精分波段 

分 类 号：P209[天文地球—测绘科学与技术]