Object-Based Downscaling Method for Land Surface Temperature with High-Spatial-Resolution Multispectral Data DOI Creative Commons
Shibao Wu, Shengmao Zhang, Fei Wang

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(8), P. 4211 - 4211

Published: April 11, 2025

Land surface temperature (LST) is an important environmental parameter in many fields. However, studies require high-spatial- and high-temporal-resolution LST products to improve the coarse spatial resolution of moderate-resolution imaging spectroradiometer (MODIS) LSTs. Numerous approaches have downscaled MODIS images a finer using pixel-based image analysis (PBA). Meanwhile, object-based (OBIA) methods, which developed rapidly high-spatial-resolution visible near-infrared (VNIR) band data, received little attention downscaling field. In this paper, we propose (OBD) method for multispectral (e.g., Landsat Thematic Mapper (TM), Enhanced Plus (ETM+), Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER)) as auxiliary data. The fundamental principle preserve thermal radiance “object”, composed several pixels (partly or entirely) unchanged after disaggregation into subpixels resulting image. decomposition process consists two key parts: (TR) estimation object from weight calculation sub-objects subpixels. Objects were generated VNIR data remote sensing indices normalized difference vegetation index (NDVI), built-up (NDBI), fractions different endmembers) multiscale segmentation method. was calculated based on weights their parent objects, estimated by relationships between LST. accuracy efficiency OBD validated pair ASTER datapoints that acquired at same time. decomposed results showed distribution closely resembled true ASTER, with RMSE 2.5 K entire A comparison PBA methods pixel also indicated achieves lowest root mean square error (RMSE) across landcovers, including urban areas, water bodies, natural terrain. Therefore, proposed significantly enhances capability increasing LST, providing alternative improving expanding applicability high-temporal- products.

Language: Английский

Object-Based Downscaling Method for Land Surface Temperature with High-Spatial-Resolution Multispectral Data DOI Creative Commons
Shibao Wu, Shengmao Zhang, Fei Wang

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(8), P. 4211 - 4211

Published: April 11, 2025

Land surface temperature (LST) is an important environmental parameter in many fields. However, studies require high-spatial- and high-temporal-resolution LST products to improve the coarse spatial resolution of moderate-resolution imaging spectroradiometer (MODIS) LSTs. Numerous approaches have downscaled MODIS images a finer using pixel-based image analysis (PBA). Meanwhile, object-based (OBIA) methods, which developed rapidly high-spatial-resolution visible near-infrared (VNIR) band data, received little attention downscaling field. In this paper, we propose (OBD) method for multispectral (e.g., Landsat Thematic Mapper (TM), Enhanced Plus (ETM+), Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER)) as auxiliary data. The fundamental principle preserve thermal radiance “object”, composed several pixels (partly or entirely) unchanged after disaggregation into subpixels resulting image. decomposition process consists two key parts: (TR) estimation object from weight calculation sub-objects subpixels. Objects were generated VNIR data remote sensing indices normalized difference vegetation index (NDVI), built-up (NDBI), fractions different endmembers) multiscale segmentation method. was calculated based on weights their parent objects, estimated by relationships between LST. accuracy efficiency OBD validated pair ASTER datapoints that acquired at same time. decomposed results showed distribution closely resembled true ASTER, with RMSE 2.5 K entire A comparison PBA methods pixel also indicated achieves lowest root mean square error (RMSE) across landcovers, including urban areas, water bodies, natural terrain. Therefore, proposed significantly enhances capability increasing LST, providing alternative improving expanding applicability high-temporal- products.

Language: Английский

Citations

0