基于二次多项式的AOD数据融合方法

AOD Data Fusion Method Based on Quadratic Polynomial

针对Aqua和Terra MODIS AOD数据利用线性回归算法拟合结果不够精确的问题,本文提出了二次多项式回归算法对其进行拟合,二次多项式是指这个多项式的项数超过1,且最高次方数为2。采用二次多项式回归和线性回归算法分别对2015年随机选择的一天和4—6月的AOD数据进行拟合,并将两种方法拟合的结果进行对比分析。研究结果显示,针对同一组Aqua和Terra MODIS AOD数据的拟合,二次多项式回归方法拟合得到的RMSE、MAE、R值比线性回归拟合方法得到的值精度都要高很多,说明二次多项式回归拟合方法在Aqua和Terra MODIS AOD数据的拟合方面优于线性回归方法的拟合,证明了二次多项式回归拟合方法适用于此方面的研究,而且能够提升Aqua和Terra MODIS AOD数据拟合结果的精度。

Aiming at the problem that Aqua and Terra MODIS AOD data are not accurate enough by linear regression algorithm, this paper proposes a quadratic polynomial regression algorithm to fit it, the quadratic polynomial means that the polynomial has more than 1 number and the highest square number is 2. The quadratic polynomial regression and linear regression algorithm were used to fit the randomly selected day and the AOD data from April to June respectively, and the results of the two methods are compared and analyzed. The results show that the RMSE, MAE,R values obtained by the quadratic polynomial regression method are much better than those obtained by the linear regression fitting method for the fitting of the same group of Aqua and Terra MODIS AOD data.It is shown that the quadratic polynomial regression fitting method is better than the fitting of the linear regression method in the fitting of Aqua and Terra MODIS AOD data, and it is proved that the quadratic polynomial regression fitting method is suitable for this research and can improve the accuracy of Aqua and Terra MODIS AOD data fitting results.