VSDF: A variation-based spatiotemporal data fusion method

Xu, Chen; Du, Xiaoping; Yan, Zhenzhen; Zhu, Junjie; Xu, Shengyu; Fan, Xiangtao

doi:10.1016/j.rse.2022.113309

Cited by 16 publications

(4 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LOTSFM is structured with reference to Fit-FC, a spatiotemporal fusion model that explicitly decomposes the structure into regression model fitting (RM), spatial filtering (SF), and residual compensation (RC) [66]. Many spatiotemporal fusion models can identify components similar to those of Fit-FC [67][68][69]. Additionally, the structure of the LOTSFM was adjusted to address the problem of long-time-series fusion for LST.…”

Section: Fusion Model Structurementioning

confidence: 99%

“…Accuracy is assessed with reference to the all-round performance assessment (APA) [81], which provides a comprehensive and standard model assessment framework adopted by new spatiotemporal fusion models, such as variation-based spatiotemporal data fusion (VSDF) [69] and the comprehensive flexible spatiotemporal data fusion (CFSDAF) [63]. Specifically, the APA proposes using metrics including the average difference (AD), rootmean-square error (RMSE), Robert's edge (EDGE), and local binary patterns (LBP) as evaluation criteria for assessing the performance of spatiotemporal fusion algorithms.…”

Section: Regional Replicability Assessment Of Lotsfmmentioning

confidence: 99%

See 1 more Smart Citation

A Spatiotemporal Fusion Model of Land Surface Temperature Based on Pixel Long Time-Series Regression: Expanding Inputs for Efficient Generation of Robust Fused Results

Chen,

Zhang,

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Spatiotemporal fusion technology effectively improves the spatial and temporal resolution of remote sensing data by fusing data from different sources. Based on the strong time-series correlation of pixels at different scales (average Pearson correlation coefficients > 0.95), a new long time-series spatiotemporal fusion model (LOTSFM) is proposed for land surface temperature data. The model is distinguished by the following attributes: it employs an extended input framework to sidestep selection biases and enhance result stability while also integrating Julian Day for estimating sensor difference term variations at each pixel location. From 2013 to 2022, 79 pairs of Landsat8/9 and MODIS images were collected as extended inputs. Multiple rounds of cross-validation were conducted in Beijing, Shanghai, and Guangzhou with an all-round performance assessment (APA), and the average root-mean-square error (RMSE) was 1.60 °C, 2.16 °C and 1.71 °C, respectively, which proved the regional versatility of LOTSFM. The validity of the sensor difference estimation based on Julian days was verified, and the RMSE accuracy significantly improved (p < 0.05). The accuracy and time consumption of five different fusion models were compared, which proved that LOTSFM has stable accuracy performance and a fast fusion process. Therefore, LOTSFM can provide higher spatiotemporal resolution (30 m) land surface temperature research data for the evolution of urban thermal environments and has great application potential in monitoring anthropogenic heat pollution and extreme thermal phenomena.

show abstract

Section: Fusion Model Structurementioning

confidence: 99%

Section: Regional Replicability Assessment Of Lotsfmmentioning

confidence: 99%

A Spatiotemporal Fusion Model of Land Surface Temperature Based on Pixel Long Time-Series Regression: Expanding Inputs for Efficient Generation of Robust Fused Results

Chen,

Zhang,

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…They are suitable for capturing changes such as phenological and land cover changes. Typical models include FSDAF [45], FSDAF 2.0 [46], TC-Umixing [47], RASDF [48], and VSDF [49].…”

Section: Introductionmentioning

confidence: 99%

The Improved U-STFM: A Deep Learning-Based Nonlinear Spatial-Temporal Fusion Model for Land Surface Temperature Downscaling

Guo,

Li,

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

The thermal band of a satellite platform enables the measurement of land surface temperature (LST), which captures the spatial-temporal distribution of energy exchange between the Earth and the atmosphere. LST plays a critical role in simulation models, enhancing our understanding of physical and biochemical processes in nature. However, the limitations in swath width and orbit altitude prevent a single sensor from providing LST data with both high spatial and high temporal resolution. To tackle this challenge, the unmixing-based spatiotemporal fusion model (STFM) offers a promising solution by integrating data from multiple sensors. In these models, the surface reflectance is decomposed from coarse pixels to fine pixels using the linear unmixing function combined with fractional coverage. However, when downsizing LST through STFM, the linear mixing hypothesis fails to adequately represent the nonlinear energy mixing process of LST. Additionally, the original weighting function is sensitive to noise, leading to unreliable predictions of the final LST due to small errors in the unmixing function. To overcome these issues, we selected the U-STFM as the baseline model and introduced an updated version called the nonlinear U-STFM. This new model incorporates two deep learning components: the Dynamic Net (DyNet) and the Chang Ratio Net (RatioNet). The utilization of these components enables easy training with a small dataset while maintaining a high generalization capability over time. The MODIS Terra daytime LST products were employed to downscale from 1000 m to 30 m, in comparison with the Landsat 7 LST products. Our results demonstrate that the new model surpasses STARFM, ESTARFM, and the original U-STFM in terms of prediction accuracy and anti-noise capability. To further enhance other STFMs, these two deep-learning components can replace the linear unmixing and weighting functions with minor modifications. As a deep learning-based model, it can be pretrained and deployed for online prediction.

show abstract

“…STARFM is a weight functionbased approach, FSDAF is a hybrid method, and Fit-FC is one of the few methods designed to combine S2 and S3 OLCI images. Although all these methods have thoroughly considered the spatial and temporal differences of the input images, spectral discrepancies have received less attention [14]. This limitation hinders their application in studies of heterogeneous landscapes [15].…”

mentioning

confidence: 99%

Adjustment of Sentinel-3 Spectral Bands With Sentinel-2 to Enhance the Quality of Spatio-Temporally Fused Images

Boumahdi,

García-Pedrero,

Lillo-Saavedra

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Spatiotemporal fusion (STF) methods are a 1 paramount solution for generating high spatial and temporal 2 time series, overcoming the limitations of spatial and temporal 3 resolution of satellite data. STF methods typically rely on band-4 by-band fusion, assuming spectral similarities. However, selecting 5 the optimal band for fusion becomes challenging when multiple 6 narrow bands overlap with the target band, often leading to the 7 use of only one single band. Furthermore, sensor specifications 8 and observation configurations can further compound this chal-9 lenge, reducing spectral and spatial information. 10 We introduce a new preprocessing step that maximizes the 11 use of spectral information from narrow bands. It minimizes 12 radiometric differences caused by sensor variations in the STF 13 process by considering the spectral response function (SRF). 14 Our method generates adjusted bands that closely match the 15 target band's spectral characteristics, leveraging all available 16 spectral information. We evaluated this strategy at two study 17 sites employing Sentinel 2 and Sentinel 3 data by comparing 18 fused images from adjusted bands and the original bands using 19 three popular STF methods. 20 The results obtained showed that the images fused with the 21 adjusted bands were closer to the target images and achieved 22 better performance, improving the fusion quality compared to 23 the original bands (SAM by 37% and RMSE by 30%). The 24 preprocessing step offers a feasible approach to generate spectral 25 bands that would be captured by the sensors if they had the 26 same spectral characteristics. Importantly, this preprocessing 27 technique is applicable to any STF method. 28 Index Terms-Spatiotemporal data fusion, Sentinel-2, Sentinel-29 3 OLCI, Spectral Response Function, Bands overlapping, Band 30 adjustment 31 I. INTRODUCTION 32 A S sensor technologies have advanced, there has been 33 a noticeable improvement in the spatial, spectral, and 34 temporal resolutions of images captured by optical remote 35 sensing sensors used for land cover studies. However, technical 36 limitations create a fundamental trade-off among these reso-37 lutions [1]. For instance, the MultiSpectral Instrument (MSI) 38 on board Sentinel-2 (S2) provides images with high spatial 39 resolutions (10 m, 20 m, and 60 m) and a frequency of re-40 visiting of 5 days. On the other hand, the optical instruments 41

show abstract

VSDF: A variation-based spatiotemporal data fusion method

Cited by 16 publications

References 40 publications

A Spatiotemporal Fusion Model of Land Surface Temperature Based on Pixel Long Time-Series Regression: Expanding Inputs for Efficient Generation of Robust Fused Results

A Spatiotemporal Fusion Model of Land Surface Temperature Based on Pixel Long Time-Series Regression: Expanding Inputs for Efficient Generation of Robust Fused Results

The Improved U-STFM: A Deep Learning-Based Nonlinear Spatial-Temporal Fusion Model for Land Surface Temperature Downscaling

Adjustment of Sentinel-3 Spectral Bands With Sentinel-2 to Enhance the Quality of Spatio-Temporally Fused Images

Contact Info

Product

Resources

About