定位定向系统数据的航空高光谱影像几何校正

Geometric correction of airborne HySpex hyperspectral image based on POS data

针对航空高光谱数据几何畸变严重的问题,该文探索了利用POS数据辅助航空高光谱影像几何校正的研究现状,分析了基于POS系统的导航数据的线阵推扫高光谱数据解算成像瞬间外方位元素的原理,以及行方位数据支持的线中心投影影像通过改进的共线方程方法几何校正的过程,实现了无地面控制点数据的直接几何校正。采用高光谱影像分块处理方法,实现了高吞吐快速几何校正。HySpex SWIR-384高光谱影像的实验结果表明,该文的校正算法可达到9.15m的绝对定位精度,最终利用少量地面控制点进行几何精校正,其均方根误差达到0.93m。

Research on the progress of the geometric correction, airborne hyperspectral imagery with POS data was explored, and the theory of linear array push broom of hyperspectral data imaging was combined with instantaneous exterior orientation elements based on the navigation data of POS system. An improved method of geometric correction of the collinear equation based on the center line data of the projected image position data was presented. High throughput and fast geometric correction was achieved through image block processing method. Lastly, the direct geometric calibration of the data was implemented with no ground control points. Experimental results of HySpex SWIR-384 hyperspectral images indicated that the absolute positioning accuracy of the algorithm, using a small amount of ground control points, was 9.15 m and the RMSE was 0.93 m.