CAJ | 학술논문

针对无人机在区域范围拍摄的低空遥感影像数量多，且相邻影像间存在旋转、尺度变化大等导致的立体定向困难问题，本文提出一种适用于无人机影像自动相对定向及模型连接的流程。即利用SURF算法对区域无人机影像进行特征提取后，连续相对定向及模型连接过程使用结合相对定向模型的RANSAC算法去除错匹配点，并针对RANSAC算法选取样本匹配点对容易陷入局部最优问题进行改进，提高了结果精度。
침대무인궤재구역범위박섭적저공요감영상수량다，차상린영상간존재선전、척도변화대등도치적입체정향곤난문제，본문제출일충괄용우무인궤영상자동상대정향급모형련접적류정。즉이용SURF산법대구역무인궤영상진행특정제취후，련속상대정향급모형련접과정사용결합상대정향모형적RANSAC산법거제착필배점，병침대RANSAC산법선취양본필배점대용역함입국부최우문제진행개진，제고료결과정도。
According to the number of low-altitude UAV remote sensing images always large and image distortion as a result of the rotation and scaling difference between neighboring images, we propose a low-altitude UAV remote sensing images suitable for high-precision automatic relative orientation and models connection.In this paper, SURF algorithm is used to extract multiple low-altitude remote sensing image feature points, then during relative orientation adoption and models connection we use the RANSAC algorithm to remove the wrong match points, and improve the problem of RANSAC algorithm which is easy to fall into local optimization, which can improve the results accuracy.