Satellite Image Time Series Decomposition Based on EEMD

Abstract

Satellite Image Time Series (SITS) have recently been of great interest due to the emerging remote sensing capabilities for Earth observation. Trend and seasonal components are two crucial elements of SITS. In this paper, a novel framework of SITS decomposition based on Ensemble Empirical Mode Decomposition (EEMD) is proposed. EEMD is achieved by sifting an ensemble of adaptive orthogonal components called Intrinsic Mode Functions (IMFs). EEMD is noise-assisted and overcomes the drawback of mode mixing in conventional Empirical Mode Decomposition (EMD). Inspired by these advantages, the aim of this work is to employ EEMD to decompose SITS into IMFs and to choose relevant IMFs for the separation of seasonal and trend components. In a series of simulations, IMFs extracted by EEMD achieved a clear representation with physical meaning. The experimental results of 16-day compositions of Moderate Resolution Imaging Spectroradiometer (MODIS), Normalized Difference Vegetation Index (NDVI), and Global Environment Monitoring Index (GEMI) time series with disturbance illustrated the effectiveness and stability of the proposed approach to monitoring tasks, such as applications for the detection of abrupt changes.

Trend Analysis of Relationship between Primary Productivity, Precipitation and Temperature in Inner Mongolia

“…As a self-adaptive approach, EMD decomposes the non-stationary signal based on its own peculiarity [48]. Kong et al [39] presented a rigorous decomposition approach for MODIS NDVI with ensemble EMD. In his study, the trend is generated by the combination of lower frequency components decomposed from the original data and the residual which is an overall trend of time series [49].…”

“…Wu and Huang [50] claimed that the overall trend that is derived as an intrinsically fitted monotonic function where there can be at most one extremum within a given span. Therefore, in this study, we adopted the residual, the trend of overall time series [39], to analyze vegetation response to climate change.…”

“…The data used in this paper include three datasets: Modis EVI (enhanced vegetation index), meteorological observations (of precipitation and temperature), and IMP land cover map 2010. EVI was adopted as a proxy of the vegetation primary productivity instead of NDVI [35,39,40], since EVI overcomes the saturation and soil noise problems of NDVI by decreasing the atmospheric effect and reducing the canopy background [41]. The 16-day EVI data with 250 m spatial resolution for 2000-2014 were collected from MODIS 13Q1, which is available at https://lpdaac.usgs.gov/data_access/data_ pool.…”

The Earth's natural pulse electromagnetic fields for earthquake time–frequency characteristics: Insights from the EEMD–WVD method

“…To solve this problem, Ensemble Empirical Mode Decomposition (EEMD), which can indeed improve IMFs' unclearness, has been proposed. The main idea of the EEMD method is the use of statistical characteristics with frequencies evenly distributed with added Gaussian white noise, allowing the signal to reach different scales of continuity, and reducing the degree of modal aliasing (Kim & Cho, 2016;Kong, Meng, Li, Yue, & Yuan, 2015). The EEMD process of the signal S(t) is shown in Figure 4 ( Hawinkel et al, 2015;Wu, Huang, Wallace, Smoliak, & Chen, 2011).…”

Comparative study of three satellite image time-series decomposition methods for vegetation change detection

“…Spectral-frequency approaches allow the characterization of the most dominant frequency variations within a SITS, for instance using harmonic representation (Jakubauskas, Legates, & Kastens, 2001). Wellknown approaches are the (fast) Fourier transform (Azzali & Menenti, 2000), multi-resolution analysiswavelet transform (MRA-WT) (Martínez & Gilabert, 2009;Percival, Wang, & Overland, 2004), empirical mode decomposition (Kong, Meng, Li, Yue, & Yuan, 2015) and singular spectrum analysis (Mahecha, Fürst, Gobron, & Lange, 2010). In contrast, statistical approaches aim to decompose the time series into the previously mentioned types of land-surface dynamics, i.e.…”