High-quality vegetation index product generation: A review of NDVI time series reconstruction techniques

•A review on high-quality NDVI time series product generation.•Three kinds of methods are summarized.•The merits and demerits of each technique are introduced.•Five potential development trends are revealed. Normalized difference vegetation index (NDVI) derived from satellites has been ubiquitously utilized in the field of remote sensing. Nevertheless, there are multitudinous contaminations in NDVI time series because of the atmospheric disturbance, cloud cover, sensor failure, and so on. It is crucial to remove the noises prior to further applications. Numerous techniques have been proposed to alleviate this issue in the last few decades. To the best of our knowledge, there hasn’t been a systematical study to summarize and analyze the status of NDVI time series reconstruction techniques since 1980s. As a result, our goal is to recapitulate the current approaches for reconstructing high-quality NDVI time series, followed by an interpretation on the principle, merits and demerits of different kinds of methods. They were mainly classified into temporal-based methods, frequency-based methods and hybrid methods. The evaluation approaches on the quality of NDVI reconstruction were introduced, accompanied with the future development tendency.