Adaptive Reconstruction of Time Series Satellite Data Based on Multi-Periodic Harmonic Analysis

Satellite data have been widely used in many applications such as monitoring changes of land-cover. Remotely sensed data inevitably contain disturbances caused by cloud presence, atmospheric variability and instrument problems, which limit the application utility. Noise reduction and reconstruction of high-quality data streams is important for the accuracy and reliability of the data analysis. This study proposes an adaptive reconstruction system for time series satellite data using multi-period harmonic analysis. A complex model of multiple periods would be proper to embody inter-annual and inner-annual variations of surface parameters. The experimental results with both simulation data and the remotely sensed data show the potentiality of the proposed adaptive reconstruction system. In addition, it has an advantage of real-time data reconstruction by adaptive estimation approach.