GF-1遥感大数据自动化正射校正系统设计与实现

发布时间：2018-04-19 15:38

本文选题：遥感 + 数据处理　；参考：《农业工程学报》2017年S1期

【摘要】：近年来随着遥感数据的爆炸性增长,快速、稳定的自动化影像正射校正成为遥感大数据处理的重要环节。该文在分析GF-1遥感大数据组织方式与元数据特征的基础上,将有理多项式模型正反变换与数字高程数据提取结合,设计实现自动化正射校正系统,并以提高正射校正计算效率与稳定性为目标,研究待校正影像对应数字高程数据快速提取方法,待校正影像分块读取策略等关键问题。在此基础上针对20景覆盖不同地形区域GF-1 8 m多光谱正射校正影像选择均匀分布的检查点,以Google Earth影像中同名点坐标为真值,分析校正误差及收敛情况,试验结果 X(纬线方向)方向和Y(经线方向)方向最大误差均小于16.863 m,距离误差小于23 m,并且92.25%的检查点误差小于16 m(2个像元)。该文提出的自动化正射校正方案在山地地形与平原地形均表现出良好的校正精度与稳定性。
[Abstract]:In recent years, with the explosive growth of remote sensing data, rapid and stable automatic orthophoto correction has become an important link in remote sensing big data processing.Based on the analysis of GF-1 remote sensing big data organization mode and metadata features, this paper combines rational polynomial model forward and inverse transformation with digital elevation data extraction to design and implement an automatic forward correction system.In order to improve the efficiency and stability of orthophoto correction, the key problems such as fast extraction method of digital elevation data and reading strategy of block reading for corrected image are studied in order to improve the efficiency and stability of orthophoto correction.On the basis of this, a uniform distribution checkpoint is selected for the GF-1 8 m multispectral orthophoto correction image covering 20 different terrain areas. The coordinate of the same name point in the Google Earth image is taken as the true value, and the correction error and convergence are analyzed.The results show that the maximum errors of direction X (latitude direction) and direction Y (longitude direction) are less than 16.863 m, distance error is less than 23 m, and the error of 92.25% check point is less than 16 m (2 pixels).The automatic forward correction scheme presented in this paper shows good accuracy and stability in both mountainous terrain and plain terrain.
【作者单位】：中国农业大学农业部农业信息获取技术重点实验室;中国科学院遥感与数字地球研究所国家遥感应用工程技术研究中心;中国农业大学国土资源部农用地质量与监控重点实验室;
【基金】：国家自然科学基金资助项目(41171337)
【分类号】：P237

【相似文献】