摘要
无人机数字影像配准技术是无人机航空摄影测量的基础技术之一。针对无人机遥感影像配准问题,本文选取地物特征点进行坐标系投影变换,从而确定影像空间位置,完成影像配准工作,并进行精度评定。研究表明:①通过手动选取控制点进行坐标拟合,可有效将TIF和JPEG影像进行配准。②基于灰度共生矩阵的特征点匹配法具有较高的配准精度,基于相位相关的区域配准法适用于处理大面积的影像配准问题。③利用TIF影像数据配准精度较高,JPEG影像数据更便于传输与存储。④配准的最佳方法取决于所应用的场景。
Digital image registration technology of unmanned aerial vehicles(UAVs)is one of the basic technologies of UAV aerial photogrammetry.In view of the remote sensing image registration problem of UAVs,this paper selected the feature points of ground objects to carry out the coordinate system projection transformation,so as to determine the spatial position of the image,achieve image registration,and evaluate the precision.The results show that:①TIF and JPEG images can be registered effectively by manually selecting control points for coordinate fitting.②The feature point matching method based on the gray co-occurrence matrix has high registration accuracy,and the region registration method based on phase correlation is suitable for large-area image registration.③The registration precision of TIF image data is higher,and JPEG image data is more appropriate for transmission and storage.④The best method of registration depends on the application scenario.
作者
钟镇声
高崇越
ZHONG Zhensheng;GAO Chongyue(Guangdong Fifth Dimension Technology Company Limited,Guangzhou Guangdong 510663,China;Department of Geological Engineering,Qinghai University,Xining Qinghai 810016,China)
出处
《北京测绘》
2023年第12期1607-1611,共5页
Beijing Surveying and Mapping
基金
青海省重点研发与转化计划项目(2023-SF-122)。
关键词
无人机遥感影像
影像配准
重采样
Matlab拟合
精度分析
unmanned aerial vehicle(UAV)remote sensing image
image registration
resampling
Matlab fitting
precision analysis
