Change detection of MODIS time series using a wavelet transform

This paper presents a rapid and easy-to-use approach for change detection of NDVI time series using wavelet transform. Wavelet transform has a long history in signal and image processing field. However, the research on the large-scale remote sensing images by wavelet transform is rarely. The method in this paper is practical by using wavelet analysis on the large scale remote sensing time series. The time series based on Normalized Difference Vegetation Index (NDVI) are fundamental to analyze the dynamic nature of the vegetation. The trend of NDVI time series is extracted by using wavelet transform. For more accurate results, a seeded region growing algorithm is used for the detailed study on the certain areas. Results from the Tibetan Plateau show the wavelet transform in combination with the region growing procedure provides a efficient approach to estimate the areas where vegetation changed.