Abstract The Gaia mission’s third data release (Gaia DR3) offers extensive observations of small solar system objects, presenting a key opportunity to expand the asteroid property database. Taxonomy, diameter, and albedo are fundamental physical parameters for characterizing asteroids. In the absence of thermal infrared observations, diameter estimates primarily rely on the Bowell relationship, with reliability dominated by geometric albedo and absolute magnitude. To improve inversion accuracy using optical data and to obtain parameters for a larger number of asteroids, we propose an asteroid albedo-diameter joint inversion method (AadRF) that integrates low-resolution spectra, orbital dynamics, and physical parameters through artificial intelligence (AI). Cross validation and independent testing show that AadRF reduces the root mean square error for albedo and diameter predictions by 64.0 percentage points and 70.2 percentage points, respectively, compared to the traditional method, with corresponding mean absolute percentage errors of 28.9% and 17.5%. The model’s advantage lies in its fusion of multisource information, embedded domain knowledge, and a hybrid deep learning–random forest architecture, which together enhance generalization. Applied to Gaia DR3, the proposed methods yield a catalog of taxonomic types, albedos, and diameters for 58,168 asteroids. Compared to existing databases, it expands the number of spectrally classified samples by nearly tenfold and adds approximately 18,100 new entries for both albedo and diameter. Further statistics reveal distinct patterns in size distribution, albedo clustering, and taxonomic composition across orbital populations. This study demonstrates the potential of AI-driven, multisource fusion for estimating asteroid parameters from large data sets and broader applications in small body property inversion.
葛 et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: