【摘要】 MODIS影像是一种新型和重要的数据。对MODIS 1B影像几何畸变原因进行了深入分析 ,选择了一种合适的纠正方法。对于 1km分辨率MODIS 1B影像 ,直接采用 1km分辨率的空间坐标进行几何纠正 ;对于2 5 0m和 5 0 0m分辨率的MODIS 1B影像 ,先用三次样条曲线对坐标进行插值生成同分辨率的坐标 ,然后利用坐标插值结果对其进行几何纠正。由于MODIS影像在空间分布上的特殊性 ,采用前向和后向映射相结合的方式确定纠正后某一像素点在原始影像中的位置。根据该位置的条带重叠度 ,可以确定参与计算像素个数、搜索窗口的大小以及窗口的精确位置 ;采用归一化反距离加权插值法计算纠正后像素点的属性值。上述优化算法不仅保证了纠正后影像的质量 ,而且提高了数据处理速度。作者在VisualC ++6 0环境下开发实现了上述算法。从坐标插值和几何纠正结果分析 ,无论是数据处理速度还是纠正后的影像质量和精度均达到要求更多还原

【Abstract】 MODIS image is a kind of new and important data for global change research. After analyzing in the depth causes of MODIS data distortion, an optimal method to remove the geometric distortion is developed. The longitude and latitude coordinates, obtained by MODIS sensor, are used to remove the distortion of the 1km resolution MODIS data; For 250m and 500m resolution MODIS data, the same resolution coordinates are calculated by the spline interpolation, then the geometric distortion is removed. Because of the spatial irregularity of MODIS 1B data, a method with forward gridding approach and the inverse gridding approach is utilized to allocate a pixel’s position of the geometrically corrected data in the original image. By computing the overlap degree of the current position, the number of the pixels used to calculate the value of the output pixels, the search window sizes and the accurate position of the windows are decided. The value of output pixels is calculated by the distance inverse weight method. All optimal approaches above will improve the processed image’s quality and the processing efficiency. The software based on the above approaches is developed in the Visual C ++6.0 environment. From the result, the conclusion that the approaches are practical and reasonable can be made.更多还原