基于时空相关性的NDVI时序重建方法研究

Study on the reconstruction of NDVI time series based on spatial-temporal correlation

针对目前时序NDVI数据重建算法时空相关性考虑不足的问题,提出一种综合利用时间域Savitzky-Golay滤波、空间域平滑滤波和异常数据剔除算法的NDVI时序重建方法.并以10年的SPOT VGT NDVI S10产品时序影像为实验数据,对算法进行了测试.结果表明,该方法对数据集中的噪声具有很好的去除效果,可以为基于NDVI的定量分析研究提供高质量数据.

Aiming to solve the problem that most present reconstruction algorithms of NDVI time series have not considered the spatial-temporal correlation,this paper proposed a new method to reconstruct NDVI time series by comprehensively combining Savitzky Golay filter,spatial smoothing filter and anomaly removal algorithm.The method is tested by using ten-years time-series data of SPOT,VGT,NDVI,S10.The results show that this method proposed can effectively remove noises in the long timeseries dataset and can provide high quality data for quantitative analysis base on NDVI.