朱思源（毕业论文）终稿.docVIP

PAGE \* MERGEFORMAT 17 MODIS 1B影像的Bow-Tie效应快速去除技术 Rapid Removal of Bow-Tie Effects from MODIS 1B Data: An Algorithmic Development 目 录 摘要3 Abstract4 引言5 算法设计6 Bow-Tie现象成因分析6 MODIS 1B的纬度科学数据集7 Bow-Tie去除算法流程8 算法实现12 纬度数据的蝴蝶结现象去除代码示例12 图像数据的蝴蝶结现象去除代码示例14 结果与分析15 结论与讨论17 参考文献19 致谢19 附录1 程序运行界面20 附录2 程序源代码21 MODIS 1B影像的Bow-Tie效应快速去除技术 摘 要：中分辨率成像光谱仪（MODIS）是一款广泛应用于进行大、中尺度大气、海洋和陆地综合监测的新一代被动式摆动扫描遥感传感器，其±55o的宽视场角和2330公里的幅宽使其影像边缘区域存在明显的重叠状几何畸变，称为Bow-Tie效应或蝴蝶结效应。目前该类几何畸变通常在进行投影转换时同步去除，但投影转换处理耗时较长，不利于影像的直接快速浏览。而现有Bow-tie效应去除算法除了ENVI外多直接基于影像上重叠区的检测，未充分利用MODIS自带的经纬度信息，算法效率不高。本文从MODIS影像的成像机制及其1B文件的存储结构入手，根据其自带的星下点1公里间隔纬度数据计算扫描方向上各列的重叠程度，并以此为依据进行Bow-tie效应去除。此外，鉴于部分纬度数据可能由于仪器等原因出现异常值，在本文的算法中也对这些异常进行了纠正。本算法在Visual C++ 6.0环境下开发实现，数据处理速度与纠正后影像质量均达到了设计要求。 关键词：MODIS; Bow-Tie效应去除; 算法设计; 畸变去除 Rapid Removal of Bow-Tie Effects from MODIS 1B Data: An Algorithmic Development Abstract: Moderate Resolution Imaging Spectroradiometer (MODIS) is a passive swing spectral scanner of new generation for collecting remote sensing data to be used in building a global dynamic model of the earth environment. Because of its ±55o wide observation angles, obvious overlapping, known as the Bow-Tie effect, occurs in the edge area of MODIS scanned images, which are provided to end users as the 1B data. The existing algorithms for the removal of the Bow-Tie effect usually integrate this process into the image projection, which takes a considerable amount of time before seeing the results. In addition most of these algorithms are based on direct detection of the overlap between adjacent swaps and do not utilize the relevant geolocation information contained in the HDF format. The study presented here attempts to develop a computer algorithm by focusing on the imaging mechanism of MODIS and the data structure of the 1B file. In this algorithm, the Bow-Tie effect is removed according to the overlap of its 1KM latitude coordinate values. Furthermore, this algorithm also provides functions for missing data exam