摘要
针对大高差困难地区高效率建立三维地理信息模型,以某山区县城为研究对象,使用旋翼和固定翼无人机不同高差航线融合为例,提出一种基于影像重叠度的大高差航带公共接边融合的方法,将多平台摄影的影像序列约束匹配,使公共接边区域检查点的匹配精度进一步提高。对影像重建匹配方法进行探讨,提出影像融合时,利用重叠区域的控制点先构建尺度空间金字塔主体影像,再根据空间特征构建局部细节影像的方法,消除了曝光点接近基线过短的影响。通过与RTK(中海达H32豪华版)采集的24个检查点坐标进行精度检验,证明其满足1∶1000大比例尺精度要求。该方法能够融合不同平台设备拍摄的影像,高效率完成大高差航带的区域精细化逆向建模。
This article takes a mountainous area as the research object,and uses rotor and fixed-wing drones as an example to integrate different elevation routes.Aiming at the problem of how to efficiently establish a three-dimensional geographic information model in areas with large altitude differences,a method of common edge fusion of large altitude difference air belt based on the degree of image overlap is proposed.This method constrains the matching of image sequences of multi-platform photography so that the matching accuracy of checkpoints in the common edge area is further improved.It also discusses the method of image reconstruction and matching,mentioning that in the process of image fusion,it is supposed to construct the main image of the scale-space pyramid using the control points in the overlapping area firstly,and then model the detail image according to the spatial characteristics,which eliminates the influence of the short exposure point close to the baseline.It performs the precision inspection with 24 checkpoint coordinates collected by RTK(ZHD H32 deluxe edition),which could meet the accuracy requirements of a large scale of 1∶1000.The research results can integrate the images taken by different platform equipment,and can efficiently complete the regional refined reverse modeling of the large altitude difference air belt.
作者
孙保燕
张小可
姚学杰
黄邦伟
SUN Baoyan;ZHANG Xiaoke;YAO Xuejie;HUANG Bangwei(College of Architecture and Transportation Engineering,Guilin University of Electronic,Guilin,Guangxi 541004,China)
出处
《遥感信息》
CSCD
北大核心
2021年第2期54-58,共5页
Remote Sensing Information
基金
广西创新驱动发展专项资金项目(桂科AA19182023)
广西科学研究与技术开发计划项目(桂科攻1598019-8)
广西研究生教育创新计划项目(YCSW2020164)。
关键词
尺度空间
公共接边
空间特征
重叠度
影像融合
scale space
common edge
spatial feature
overlap degree
image fusion
