New GAN reconstructs satellite land surface temperature data with high accuracy

Researchers have developed a novel Multimodal Fast Fourier Convolutional GAN designed to reconstruct land surface temperature (LST) data from satellite imagery, specifically addressing gaps caused by cloud cover. This method utilizes Fast Fourier Convolution to achieve a global receptive field and integrates data from satellite observations and synthetic aperture radar (SAR). The approach demonstrates effectiveness in recovering extensive missing regions, even in scenes with over 70% cloud-induced gaps, achieving an interquartile range of scene-averaged RMSE between 0.8 K and 1.8 K across all LST quantiles. AI

IMPACT This method could improve the accuracy and completeness of satellite-based climate and environmental monitoring.

RANK_REASON This is a research paper detailing a new model for data reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• Fast Fourier Convolutional GAN
• Landsat program
• synthetic aperture radar

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →