基于地理位置信息的图像配准方法及精度分析

发布时间：2018-03-07 21:44

本文选题：遥感　切入点：图像配准　出处：《光学学报》2017年08期 　论文类型：期刊论文

【摘要】：针对传统图像配准技术难以对海洋、沙漠、草原等特征不明显区域航空遥感图像进行配准的问题,提出了一种基于地理位置信息的航空遥感图像配准算法。依据载机定位定向系统测量的载机位置、姿态信息以及航空相机中位置编码器测量的框架角位置信息,利用齐次坐标变换的方法求解配准点在大地坐标系下的投影。利用世界大地坐标系-84坐标系定义的地球椭球模型确定匹配点的经纬度信息,将相同地理位置信息的配准点进行配准。采用蒙特卡罗法仿真分析了载机姿态位置信息及框架角位置信息对配准精度及定位精度的影响。采用实际的航空遥感图像进行实验,结果表明,在载机飞行高度低于2000m,拍摄倾斜角小于18°时,配准精度可优于3m,遥感图像中的海上控制点的定位精度优于35m。
[Abstract]:Aiming at the problem that the traditional image registration technology is difficult to register the regional remote sensing images of the sea, desert, grassland and so on, A registration algorithm of aerial remote sensing image based on geographical position information is proposed, which is based on the position and attitude information of the carrier and the frame angle position information measured by the position encoder in the aerial camera. The projection of collocation points in geodetic coordinate system is solved by using the method of homogeneous coordinate transformation. The longitude and latitude information of matching points is determined by the Earth ellipsoid model defined in the world geodetic coordinate system -84 coordinate system. The effect of attitude position information and frame angle position information on registration accuracy and positioning accuracy is simulated by Monte Carlo method. The actual aerial remote sensing image is used to carry out experiments. The results show that the registration accuracy is better than 3 m and the positioning accuracy of the offshore control points in remote sensing images is better than 35 m when the flight altitude is less than 2000m and the shooting angle is less than 18 掳.
【作者单位】：中国科学院长春光学精密机械与物理研究所;中国科学院大学;
【基金】：国家自然科学基金青年基金(61405192) 吉林省科技发展计划(20140520114JH)
【分类号】：TP751

【相似文献】