Accurate Geometric Correction for Meterogical Satellite NOAA's AVHRR Date
Accurate Geometric Correction for Meterogical Satellite NOAA's AVHRR Date
复制标题
气象卫星 NOAA AVHRR 日期的精确几何校正
DOI：
10.4287/jsprs.34.2_25
复制
发表时间：
1995
期刊：
影响因子：
--
通讯作者：
M. Takagi
中科院分区：
文献类型：
--
作者：
Yaudong Chang;M. Takagi
本文提出了一种AVHRR数据的高精度几何校正方法。我们以前的工作提出了一种利用双线性插值来加速AVHRR数据几何校正的方法，但仅利用轨道信息TBUS进行几何校正后仍存在微小的残差。GCP方法是提高几何校正精度的常用方法。通过曲线拟合的方法，以多项式形式分析了每个轨道根数和卫星报告时间（卫星内部时钟）对GCP剩余误差的趋势。引入标准正则化方法，解决了从控制点残差到各轨道根数变化的反问题。由于GCP过程在直线方向和像素方向上的残留误差是平滑变化的，因此将二维曲面样条应用于平滑约束稳定器。然而，正则化方法得到的解有时不是全局最优解，而是局部最优解。这里，分析轨道信息TBUS以决定轨道根数的校正顺序。从这个校正顺序，可以使用具有多项式表达式的算法来估计每个轨道要素的变化和卫星报告时间从GCP的残差，当局部最优解满足。最后通过实验验证了该方法的有效性，并对本文提出的几何校正方法进行了总结。
The present paper proposes an accurate geometric correction method for AVHRR data . Our previous work has proposed a method, which uses the bilinear interpolation for speeding up geometric correction of AVHRR data, but there is still tiny residual error after geometric correction only using the orbital information TBUS . GCP method is the general method for improving the accuracy of geometric correction . The tendency of each orbital element and satellite reported time (satellite internal clock) for the GCP's residual error is analyzed by curve fitting method as a polynomial expression. The standard regularization method is introduced to solve the reverse problem from the GCP's residual errors to the variation of each orbital element. The residual errors in line direction and pixel direction from GCP process are smoothly changed, therefore , two-dimensional surface spline is applied to the smooth-constrained stabilizer. However, the solution from the regularization method is not the globally optimal one but the locally optimal one sometimes. Here, the orbital information TBUS is analyzed to decide the correction order of orbital elements. From this correction order, an algorithm with the polynomial expressions can be used to estimate the variation of each orbital element and satellite reported time from the GCP's residual error, when the locally optimal solution meets. Finally, some experimental results are shown to verify the proposed method , and we will conclude our proposed geometric correction method.