Search

Drought is a complex and extremely destructive natural disaster that seriously threatens the sustainable development of human society and ecosystems. The integrated drought index with remote sensing offers an efficient approach for drought monitoring and assessment, and has become one of the main trends in drought monitoring research. The existing integrated drought indices are generally used to monitor drought conditions at the national scale, with a spatial resolution of 500 m - 1 km. However, this resolution is still too coarse for farmland, unable to capture the spatial heterogeneity of drought stress at a finer field scale (such as 30 m × 30 m), and cannot meet the needs of field-scale drought monitoring such as refined agricultural water resources management and drought loss assessment. Therefore, this paper proposes an integrated framework for monitoring vegetation drought at the field scale, which integrates spatiotemporal fusion, deep learning, remote sensing, in-situ stations, and biophysical information. First, the framework established a classification system of drought factors based on the disaster system theory, including vegetation condition, drought-pregnant environment and drought-inducing factors. Second, the dense time series of 30 m monthly vegetation conditions (greenness, moisture, temperature) from 2001 to 2014 were generated based on spatiotemporal fusion. The monthly anomalies were calculated and then integrated based on 3D Euclidean distance method to generate the 30 m Vegetation Condition Anomaly Index (VCAI). Then, the drought-pregnant environment and drought-inducing factors were integrated through deep learning to generate the 30 m Environmental Drought-Inducing Index (EDII). Finally, the joint cumulative distribution was used to couple VCAI and EDII, and the 30 m High spatial resolution Drought Response Index (HiDRI) was generated. Results showed that the comparative analysis of HiDRI with in-situ Standardized Precipitation Evapotranspiration Index (SPEI), meteorological reanalysis data and remote sensing soil moisture data performed well. The overall Pearson correlation coefficient between HiDRI and in-situ SPEI was 0.601 (p < 0.01). Meanwhile, HiDRI can effectively map 30 m × 30 m field-scale drought spatial patterns in different climate zones, including heterogeneous spatial distribution and detailed texture features. In addition, the generated results of spatiotemporal fusion and deep learning have been effectively verified. • A new integrated framework for monitoring field-scale vegetation drought. • Deep learning-based 30 m environmental drought information integration. • Spatiotemporal fusion-based 30 m vegetation conditions generation. • Effective coupling by joint cumulative distribution. • A 30 m High spatial resolution Drought Response Index (HiDRI) was generated. [ABSTRACT FROM AUTHOR]

Spatiotemporal fusion aims to improve both the spatial and temporal resolution of remote sensing images, thus facilitating time-series analysis at a fine spatial scale. However, there are several important issues that limit the application of current spatiotemporal fusion methods. First, most spatiotemporal fusion methods are based on pixel-level computation, which neglects the valuable shape information of ground objects. Moreover, many existing methods cannot accurately retrieve strong temporal changes between the available high-resolution image at base date and the predicted one. This study proposes an Object-Based Spatial Unmixing Model (OBSUM), which incorporates object-based image analysis and spatial unmixing, to overcome the two abovementioned problems. OBSUM consists of one preprocessing step and three fusion steps, i.e., object-level unmixing, object-level residual compensation, and pixel-level residual compensation. The performance of OBSUM was compared with seven representative spatiotemporal fusion methods at two agricultural sites. The experimental results demonstrated that OBSUM outperformed other methods in terms of both accuracy indices and visual effects over time-series. Furthermore, OBSUM also achieved satisfactory results in crop progress monitoring and crop mapping. Therefore, it has great potential to generate accurate and high-resolution time-series observations for supporting various remote sensing applications. • OBSUM is proposed for the spatiotemporal fusion of remote sensing images. • OBSUM considers the shape information of ground objects. • An object-level residual compensation is proposed to retrieve temporal changes. • OBSUM outperforms seven comparison methods in two agricultural sites. • OBSUM can support various agricultural remote sensing applications. [ABSTRACT FROM AUTHOR]

Abstract High-spatiotemporal-resolution land surface temperatures (LSTs) are required in various environmental applications. However, due to the trade-off between the spatial and temporal resolutions in remote sensing, such data are still unavailable. Many studies have been conducted to resolve this dilemma, but difficulties remain in generating high-spatiotemporal-resolution (i.e., diurnal, e.g., 30 m resolution) LSTs. Accordingly, this study proposes a weighted combination of kernel-driven and fusion-based methods (termed CKFM) to enhance the resolution of time series LSTs; the kernel-driven process can obtain abundant spatial details from visible bands, while the fusion-based process is applied for its spatiotemporal prediction ability. CKFM contains three parts. First, a kernel-driven method is applied to predict high-resolution LSTs via a regression relationship between simulated medium-resolution LSTs and kernels. Second, a fusion-based method is applied to predict the medium-resolution LST, the result of which is used for thin plate spline (TPS) downscaling. Finally, the results of the kernel-driven and fusion-based processes are combined via weights calculated from error estimations. Compared with existing thermal sharpening methods, CKFM has the following strengths: (1) it fully utilizes the available visible and thermal bands from multiple sensors, thereby obtaining spatial details in a variety of ways; (2) it downscales LSTs in a dynamic manner; and (3) it is suitable for heterogeneous regions. CKFM is tested with Landsat 8 and MODIS data and successfully downscales the 1 km resolution MODIS LSTs into 30 m resolution data. In both visual and quantitative evaluations, CKFM is more accurate and robust than the kernel-driven method with an improvement of 0.1–0.6 K, and it reconstructs more spatial details than the fusion-based process. Based on these characteristics, CKFM is a promising method for generating daily high spatial resolution LSTs for various environmental studies. Highlights • CKFM is a weighted combination of two traditional thermal sharpening methods. • CKFM more effectively reconstructs spatial details from the available visible bands. • CKFM is suitable for heterogeneous regions. • CKFM compensates for the deficiencies of kernel-driven and fusion-based methods. • CKFM minimizes errors by distributing predicted values via weighting. [ABSTRACT FROM AUTHOR]

Fine spatial resolution (i.e., ≤ 100 m) land surface temperature (LST) data are crucial to study heterogeneous landscapes (e.g., agricultural and urban). Some well-known spatiotemporal fusion methods like the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and the Enhanced STARFM (ESTARFM), which were originally developed to fuse surface reflectance data, may not be suitable for direct application in LST studies due to the high sub-diurnal dynamics of LST. Furthermore, the effectiveness of spatiotemporal fusion methods for LST data has not been thoroughly evaluated in previous studies that only focused on relatively small spatiotemporal extents. To address these limitations, we proposed a variant of ESTARFM, referred to as the unbiased ESTARFM (ubESTARFM), specifically designed to accommodate the high temporal dynamics of LST to generate fine-resolution LST estimates. We evaluated ubESTARFM and ESTARFM against in-situ LST and the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) LST across 12 regions throughout Australia, encompassing various land covers and environments. Independent validation showed that ubESTARFM had a bias of 2.55 K, unbiased root mean squared error (ubRMSE) of 2.57 K, and Pearson correlation coefficient (R) of 0.95 against the in-situ LST over 11,290 observations at the 12 sites, all of which were considerably better than those calculated for ESTARFM, being a bias of 4.73 K, ubRMSE of 3.80 K and R of 0.92. When compared to ECOSTRESS data, ubESTARFM LST had a bias of −1.69 K, ubRMSE of 2.00 K, and R of 0.70 over 43 near clear-sky scenes, while ESTARFM LST had a bias of 1.79 K, ubRMSE of 2.68 K, and R of 0.59. Overall, our results demonstrated that ubESTARFM can avoid systematic bias accumulation, substantially reduce uncertainty deviation, and maintain a good level of correlation with validation datasets when compared to ESTARFM. A further assessment underscored the potential of ubESTARFM for application using LST data acquired from geostationary platforms (e.g., Himawari-8), with a mean ubRMSE (R) of 2.22 K (0.97) against in-situ LST over 1327 observations at 3 sites from southeast Australia at the overpass time of MODIS/Terra. This promising method leverages reliable numeric values from coarse-resolution LST while borrowing spatial heterogeneity from fine-resolution LST and has the potential to be coupled with energy balance and/or radiative transfer models thus enabling better farm and/or regional-scale water management strategies to be implemented. Furthermore, both the input and generated LST data, encompassing a comprehensive spatial extent over diverse land covers and climatic conditions, are publicly available for benchmarking future algorithmic refinements. • We developed an unbiased variant of ESTARFM (ubESTARFM) to fuse MODIS and Landsat LST. • ubESTARFM uses coarse LST as reference to locally correct systematic bias of fine LST. • We evaluated ubESTARFM against in-situ LST and ECOSTRESS LST across Australia. • ubESTARFM reduces deviation of uncertainty while it maintains spatial details. • ubESTARFM shows potential for application on geostationary LST (e.g., Himawari-8). [ABSTRACT FROM AUTHOR]

Spatial and temporal satellite image fusion (STIF) has provided a feasible alternative for generating imagery with both high spatial and temporal resolution, thus expanding the applications of existing satellite sensors. However, a critical challenge confronting the further development of STIF is to systematically and robustly address complex land surface changes, which include land cover changes without shape changes (e.g., crop rotation) and land cover changes with shape changes (e.g., urban expansion), in addition to conventional land surface changes (e.g., phenological changes of vegetation). This paper presents the Robust Adaptive Spatial and Temporal Fusion Model (RASTFM) to tackle this challenge with one prior pair of MODIS-Landsat images. In RASTFM, land surface changes are reorganized into non-shape changes (including phenological changes and land cover changes without shape changes) and shape changes (i.e., land cover changes with shape changes), which are handled differently. However, both non-shape changes and shape changes are predicted through a Non-Local Linear Regression (NL-LR) of the subject pixel's similar neighbors. A regression based high-pass modulation is also performed as a post-processing step to improve both the spatial details and spectral fidelity of the predicted Landsat image. Unlike other STIF models (e.g., the Spatial and Temporal Adaptive Reflectance Fusion Model, STARFM), RASTFM can find similar neighboring pixels more precisely through a non-local searching strategy and derives the weights of similar neighbors more rigorously via a linear regression model. As both non-shape and shape changes are treated based on the regression of similar neighboring pixels, the land surface changes are processed in a unified manner. Experiments that use one simulated and three actual MODIS-Landsat datasets featured by different types of land surface changes were conducted to demonstrate the performance of RASTFM. Comparisons with the state-of-the-art STIF models, including weighted function, unmixing and dictionary-learning methods, show that NL-LR based RASTFM can capture the land surface changes in various landscapes more accurately and robustly in a unified manner, thereby facilitating the continuous and detailed monitoring of complex and diverse land surface dynamics. [ABSTRACT FROM AUTHOR]

Dense time-series high-resolution satellite images are extremely valuable for long-term monitoring of land dynamics. Spatiotemporal fusion (STF) techniques have been developed to integrate multi-resolution satellite images to produce data with high spatial resolution and temporal frequency. Due to the large volume and diversity of higher resolution global Earth Observation (EO) data, large-scale data processing methods need to be computationally efficient, free of parameter fine-tuning, and adaptive to various data structures without process customization. These requirements are especially critical for the production of global Seamless Data Cube (SDC). Considering the limitations of existing STF methods, we propose a ROBust OpTimization-based (ROBOT) fusion model that exploits time-series information to obtain more accurate predictions. ROBOT maintains stable under varied data conditions by adopting a temporal-coherence regularization term. And being free of parameter tuning, ROBOT can be applied to arbitrary spatiotemporally distributed data without repetitive process customization. Its performance was compared with eight representative STF methods. Results show that ROBOT outperforms existing STF methods in most cases and is computationally efficient, about 4000-fold faster than ESTARFM and 600-fold faster than FSDAF. The proposed method has demonstrated its potential for global-scale SDC generation to support subsequent remote sensing applications. • ROBOT is proposed to employ time-series data for better fusion results. • It performs stably under varied data conditions and is free of parameter tuning. • It does not require input images to be completely clean and cloud-free. • It outperforms existing algorithms in both accuracy and efficiency. • It makes the on-the-run spatiotemporal data fusion possible and thus saves storage. [ABSTRACT FROM AUTHOR]

The fusion of spatiotemporal data provides the possibility to improve both the spatial and temporal resolution of remote sensing data. Nevertheless, the performance of current spatiotemporal data fusion methods is affected by several aspects, e.g., (1) retrieval of abrupt land cover changes, (2) recovery of detailed spatial information, and (3) the need to reduce side effects related to the performance differences between sensors. Concerning the above aspects, this study proposes the use of a Variation-based Spatiotemporal Data Fusion (VSDF) method. In VSDF, an abundant variation classification (AVC) is used to identify explicitly the land cover changes for spectral unmixing. In addition, feature-level fusion is introduced to strengthen the spatial structures, i.e., the edges and texture. Furthermore, a relative reliability index (RRI) is proposed to guide the prediction process to lower the uncertainty of the input datasets. With reference to the all-round performance assessment (APA) metrics, the performance of VSDF was compared with five popular methods, i.e., Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Flexible Spatiotemporal DAta Fusion (FSDAF), FSDAF 2.0, Reliable and Adaptive Spatiotemporal Data Fusion (RASDF), and Fit-FC (regression model Fitting, spatial Filtering and residual Compensation). The experimental results demonstrated that VSDF can realize a more accurate prediction of temporal land cover changes with better spatial detail than the benchmark methods. Consequently, VSDF has the potential to generate accurate high-spatiotemporal-resolution simulations for global remote sensing studies. • VSDF was developed to fuse images with different spatiotemporal resolutions. • Relative reliability index was proposed to guide the prediction intelligently. • Abundant variation classification identified explicitly the temporal changes. • Feature-level fusion strengthened the spatial structures of the prediction. [ABSTRACT FROM AUTHOR]

Spatiotemporal data fusion, as a feasible and low-cost solution for producing time-series satellite images with both high spatial and temporal resolution, has undergone rapid development over the past two decades with more than one hundred spatiotemporal fusion methods developed. Accuracy assessment of fused images is crucial for users to select appropriate methods for real-world applications. However, commonly used assessment metrics do not comprehensively cover multiple aspects of spatiotemporal fused image quality, contain redundant information, and are not comparable across different study areas. To address these problems, this study proposed a novel framework to assess all-round performances of spatiotemporal fusion methods. Four accuracy metrics, including RMSE, AD, Edge, and local binary patterns (LBP), were selected as the optimal set of assessment metrics according to the assessment criteria. These metrics not only quantify the spectral and spatial information in the fused images but also greatly alleviate information redundancy and feature computational simplicity. Furthermore, inspired by Taylor diagrams, we designed an all-round performance assessment (APA) diagram to provide a visual tool for a comprehensive assessment of the performance of spatiotemporal fusion methods, supporting cross-comparison of different spatiotemporal fusion methods by considering the effects of input data and land surface characteristics. The case study in three typical sites demonstrated that the proposed framework can better differentiate the performances of six spatiotemporal fusion methods. This new framework can promote the cross-comparison of different spatiotemporal fusion methods and guide users to select suitable methods for real-world applications, as well as facilitate the establishment of a standard accuracy assessment procedure for spatiotemporal fusion methods. • A framework was proposed to assess all-round performances of fused images. • RMSE, AD, Edge, LBP were selected as optimal metrics. • A polar diagram was designed to visualize accuracy measurements with bottom lines. • It can guide users to select appropriate fusion methods for different applications. [ABSTRACT FROM AUTHOR]

Spatiotemporal image fusion is a potential way to resolve the constraint between the spatial and temporal resolutions of satellite images and has been developed rapidly in recent years. However, two key challenges related to fusion accuracy remain: a) reducing the uncertainty of image fusion caused by sensor differences and b) addressing strong temporal changes. To solve the above two issues, this paper presents the newly proposed Reliable and Adaptive Spatiotemporal Data Fusion (RASDF) method. In RASDF, the effects of four kinds of sensor differences on fusion are analyzed systematically. A reliability index is therefore proposed to describe the spatial distribution of the reliability in input data for image fusion. An optimization strategy based on the spatial distribution of the reliability quantified by the index is developed to improve the robustness of the fusion. In addition, an adaptive global unmixing model and an adaptive local unmixing model are constructed and utilized collaboratively to enhance the ability to retrieve strong temporal changes. The performance and robustness of RASDF were compared with six representative fusion methods for both real and simulated datasets covering both homogeneous and heterogeneous sites. Experimental results indicated that RASDF achieves a better performance and provides a more reliable image fusion solution in terms of reducing the impact of sensor differences on image fusion and retrieving strong temporal changes. • RASDF is developed to blend multi-spatiotemporal-resolution satellite images. • RASDF considers the influence of sensor differences on fusion accuracy. • A reliability index is proposed to describe the spatial distribution of the reliability in input data for image fusion. • RASDF constructs an adaptive local unmixing model to enhance the ability of retrieving strong temporal changes. • RASDF has satisfactory performance and robustness in retrieving strong phenological changes. [ABSTRACT FROM AUTHOR]

Dozens of spatiotemporal fusion methods have been developed to reconstruct vegetation index time-series data with both high spatial resolution and frequent coverage for monitoring land surface dynamics. Although several studies comparing the different fusion methods have been conducted, selecting the suitable fusion methods is still challenging, as inevitable influential factors tend to be neglected. To address this problem, this study compared six typical spatiotemporal fusion methods, including the Unmixing-Based Data Fusion (UBDF), Linear Mixing Growth Model (LMGM), Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Fit-FC (regression model Fitting, spatial Filtering and residual Compensation), One Pair Dictionary-Learning method (OPDL), and Flexible Spatiotemporal DAta Fusion (FSDAF), based on simulation experiments and theoretical analysis considering three influential factors between sensors: geometric misregistration, radiometric inconsistency, and spatial resolution ratio. The results indicate that Fit-FC achieved the best performance with the strongest tolerance to geometric misregistration when radiometric inconsistency was negligible; thus, it is the first recommended algorithm for blending normalized difference vegetation index (NDVI) imagery. Instead, the FSDAF could generate the best results if radiometric inconsistency was non-negligible. These findings could help users determine the method that is appropriate for different remote sensing datasets, and provide guidelines for developers in the future development of novel methods. • Fit-FC and FSDAF performed best in the fusion experiments with influencing factors. • Fit-FC is first recommended but relatively sensitive to radiometric inconsistency. • FSDAF is second recommended but relatively sensitive to geometric error. • Sensitivity to radiometric/geometric errors was explained by theoretical analysis. [ABSTRACT FROM AUTHOR]

Spatiotemporal fusion is a feasible solution to resolve the tradeoff between the temporal and spatial resolutions of remote sensing images. However, the development of spatiotemporal fusion algorithms has not yet reached maturity, and existing methods still face many challenges, e.g., accurately retrieving land cover changes and improving the robustness of fusion algorithms. The Flexible Spatiotemporal DAta Fusion (FSDAF) method proposed by Zhu et al. in 2016 solved the abovementioned problems to some extent. However, FSDAF has two shortcomings that can be further improved: (1) FSDAF is prone to losing spatial details and predicting a "blurrier" image due to the input of coarse pixels containing type change information and a large amount of boundary information for unmixing calculation, and (2) FSDAF does not optimize the areas of land cover change. In this paper, an improved FSDAF method incorporating change detection technology and an optimized model for changed-type areas (FSDAF 2.0) was proposed to improve the aforementioned problems. Based on the existing FSDAF algorithm, FSDAF 2.0 excludes changed pixels and boundary pixels for unmixing calculation, and establishes a model to optimize the changed pixels. Its performance was compared with that of the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), the original FSDAF, and the enhanced FSDAF that incorporates sub-pixel class fraction change information (SFSDAF). Two sites consisting of landscapes with heterogeneous and large-scale abrupt land cover changes were employed for testing. The results of the experiments demonstrate that FSDAF 2.0 effectively improves the shortcomings of FSDAF, blends synthetic fine-resolution images with higher accuracy than that of the other three methods at two different sites, and strengthens the robustness of the fusion algorithm. More importantly, FSDAF 2.0 has a powerful ability to retrieve land cover changes and provides a feasible way to improve the performance of retrieving land cover changes. Consequently, FSDAF 2.0 has great potential for monitoring complex dynamic changes in the Earth's surface. • FSDAF 2.0 is developed to fuse coarse and fine spatial resolution images. • FSDAF 2.0 integrates the change detection technology to find the areas of land cover change. • FSDAF 2.0 optimizes the unmixing process to better preserve spatial details. • FSDAF 2.0 establishes a model to optimize the values of changed-type pixels. • FSDAF 2.0 can better predict both phenological change and land cover change than original FSDAF. [ABSTRACT FROM AUTHOR]