2026-10-12T11:08:27Z https://www.scidb.cn/oai

10.57760/sciencedb.space.03705 2026-06-29T14:46:02Z

2026-06-29 Figure Data for ViSCoNet-Based High-Resolution Brightness Temperature Reconstruction in Distributed Synthetic Aperture Interferometric Radiometers doi:10.57760/sciencedb.space.03705 en This dataset contains the figure data associated with the manuscript entitled “Deep Learning-Based High Resolution Image Reconstruction for Distributed Synthetic Aperture Radiometers With Sparse Sampling.” The data were generated to support the evaluation of ViSCoNet, a physics-constrained deep learning framework for high-resolution brightness temperature reconstruction in distributed synthetic aperture interferometric radiometer systems under sparse, nonuniform, and dynamically varying visibility sampling conditions.The dataset includes reconstructed brightness temperature images, visibility-domain sampling patterns, antenna array configurations, and performance metrics used to generate the figures and tables in the manuscript. The data cover comparisons among conventional and learning-based reconstruction methods, including FFT, G-matrix, SAIR-CNN, SR-SEN, and the proposed ViSCoNet method. The included results support analyses of reconstruction accuracy, spatial resolution enhancement, robustness under amplitude-phase perturbations and missing-baseline conditions, array-configuration dependence, point-source and extended-source response, and effective radiometric sensitivity.The brightness temperature scenes used in the experiments were generated from simulated distributed SAIR observations. Realistic L-band cryosphere brightness temperature scenes derived from SMOS rSIR enhanced products and brightness-temperature-like scenes mapped from the NWPU-RESISC45 remote sensing image dataset were used as input scenes. The visibility data were synthesized according to the SAIR forward observation model, where the core array provides dense low-frequency visibility samples and distributed antenna elements provide sparse long-baseline high-frequency visibility measurements. Complex visibility data were processed using signed cube-root normalization for the real and imaginary components, and brightness temperature labels were normalized for neural network training and evaluation.This dataset can be used to reproduce and verify the figure-level result reported in the manuscript, and to support further research on sparse sampling, distributed aperture synthesis, visibility spectrum completion, and physics-constrained deep learning reconstruction for passive microwave remote sensing. passive microwave remote sensing; passive microwave brightness temperature; deep learning 任第豪 PUBLIC https://creativecommons.org/publicdomain/zero/1.0/ dataset Science Data Bank