Empirical comparison of noise reduction techniques for NDVI time-series based on a new measure |
| |
| Institution: | 1. State Key Laboratory of Remote Sensing Science, College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, PR China;2. School of Environment, Tsinghua University, Beijing 100084, PR China;3. Department of Geography, University of Utah, Salt Lake City, UT 84112, USA;1. Research Group of Hydrological and Environmental Modelling (GIHMA), Research Institute of Water and Environmental Engineering, Universitat Politècnica de València, Valencia, Spain;2. Research Group in Forest Science and Technology (Re-ForeST), Research Institute of Water and Environmental Engineering, Universitat Politècnica de València, Valencia, Spain;1. Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China;2. CSIRO Digital Productivity Flagship, Leeuwin Centre, 65 Brockway Road, Floreat Park, WA 6014, Australia;3. CSIRO Digital Productivity Flagship, GPO Box 664, Canberra, ACT 2601, Australia;1. UNSW Water Research Centre, School of Civil and Environmental Engineering, University of New South Wales, Sydney, NSW 2052, Australia;2. Department of Civil and Environmental Engineering, Clarkson University, Potsdam, NY 13699, USA;1. UNSW Water Research Centre, School of Civil and Environmental Engineering, University of New South Wales, Sydney, NSW 2052, Australia;2. Department of Environmental Sciences, University of California Riverside, Riverside, CA, 92521, USA |
| |
| Abstract: | This study empirically compared noise reduction techniques for the normalized difference vegetation index (NDVI) time-series based on a new absolute measure using a time-series of 16-day composite NDVI images extracted from the Terra Moderate Resolution Imaging Spectroradiometer (MODIS) products covering the Poyang Lake area in China. We proposed an approach to accurately extract representative NDVI temporal profiles for the 12 land cover cluster types by clustering profiles, selecting optimal number of clusters, merging and labeling clusters, and selecting the representative NDVI profiles. The geometric average of the mean average distance between the reconstructed profile and the raw profiles, and the mean average distance between the reconstructed profile and the upper envelope (Dg(nr, c)) was selected as the most appropriate measure substitutive to RMSE for the evaluation of the noise reduction effects, when the ‘true’ profiles were not available. The running median, mean value, maximum operation, end point processing, and Hanning smoothing (RMMEH) filter and iterative Savitzky–Golay filter were the two most appropriate noise reduction techniques for the NDVI temporal profiles of the study area in the evaluation of noise reduction effects by the seven techniques. The robust framework using the proposed approach for the accurate extraction of representative NDVI temporal profiles and (Dg(nr, c)) in this study, is applicable in the evaluation of noise reduction effects using different techniques and in other study areas. |
| |
| Keywords: | Vegetation index Multitemporal MODIS Land cover Monitoring |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
