姿态四元数数据辅助的卫星影像几何定位方法研究

In the mordern high-resolution remote sensing satellite, attitude quaternion as a kind of attitude data is download with the satellite images. However, we have not make full use of these attitude data efficiently,or only convert them to euler angle attitude in the existing method of geopositioning for satellite images.It is shows by research that quaternion can be used to discrible the orientation and rotation of vector,and attitude quaternion can avoid the singularity and "gimbal lock" of euler angle attitude. Based on the results of preliminary study, this project will research on the establishing and optimizating the interpolation method of exterior orientation parameters (EOPs) by using attitude quaternion, and the project will research on refining and repairing of the attitude quaternion; Then the project will research on both the method of systematic error compensation for attitude quaternion and the detection and removing method of time synchronization error for different sensors, and the project will research on the geopositioning method of high-resolution satellite imagery supported by attitude quaternion data based on the above error compensating methods. At last,considering the additional parameters model, the project will research on the self-calibration bundle block adjustment supported by attitude quaternion data when the attitude quaternion and ephemeris data are regarded as observed value with weight. Researching on the geopositioning method of high-resolution satellite imagery supported by attitude quaternion have very important theoretical applied worthiness for developing and enriching existing method of geopositioning and satisfing the needs of processing for high-resolution observation from space data in future.

在当前高分辨率测绘遥感卫星中，姿态四元数作为一种姿态数据已开始随卫星影像下传，然而在现有的卫星影像几何定位方法中，该数据没有得到充分有效利用，或者仅将其转换成欧拉角姿态进行后续处理。研究证实，四元数可以非常方便地描述空间向量的方位及旋转，姿态四元数能有效避免欧拉角姿态可能引起的奇异性和出现"方向锁定"的现象。本项目拟在前期研究的基础上，对已有的四元数外方位元素内插方法进行优化研究，然后对姿态四元数数据进行精化和漏洞修补；同时研究姿态四元数的系统误差消除技术，并研究不同传感器时间同步误差的检测和消除技术，并在此基础上研究姿态四元数数据辅助的卫星影像几何定位方法；将姿态四元数数据作为带权观测值，考虑附加参数模型，研究姿态四元数数据辅助的自检校区域网平差。通过本研究，将对拓展和丰富现有卫星影像几何定位方法，满足未来国产高分辨率对地观测数据处理有极为重要的理论和应用价值。

项目针对随卫星影像下传姿态四元数数据，深入探索了其应用潜能，借鉴现有摄影测量几何处理的理论基础，建立了一套完整的姿态四元数数据辅助的卫星影像几何处理的相关模型与算法。重点针对姿态四元数数据辅助的成像几何模型、直接定位和自检校区域网平差等关键技术提出了解决方法：(1)建立了姿态四元数数据辅助的外方位元素模型，实现了卫星线阵影像外方位元素高精度动态建模，通过构造独立参数，消除四元数参数的自相关性，构建了成像几何模型；(2)建立了利用姿态四元数数据、星历数据、影像数据直接进行几何定位的模型与算法，并改正星历姿态四元数数据和飞行方向像坐标误差，提高了直接定位的精度；(3)以姿态四元数数据、星历数据为带权观测值，设计了姿态四元数数据辅助的自检校区域网平差模型、技术方案和处理流程。研制了姿态四元数数据辅助的卫星影像几何处理软件模块，实现了姿态四元数数据辅助的卫星影像几何定位功能，通过大量的试验和实际生产单位的应用检验，解决了姿态四元数数据的直接应用问题，有效提升了定位的稳定性和可靠性，丰富了卫星遥感影像高精度定位的技术手段和方法，是遥感影像高精度几何定位处理的一条新途径。在《测绘学报》《测绘科学与工程》等期刊上发表论文8篇，出版专著1部，达到了项目预期研究目标。