Adaptive Regularized Sparse Representation for Weather Radar Echo Super-Resolution Reconstruction
Haoxuan Yuan, Qiangyu Zeng, Jianxin He · 2021 International Conference on Electronic Information Engineering and Computer Science (EIECS) · 2021
Ability of weather radar to detect extreme weather on medium and small scale is limited by the resolution of its grid data. Low-resolution weather radar data may cause false warnings and forecasts of small and medium scale in extreme weather. The aim of this work is to provide a solution for enhancing the resolution of weather radar data. A super-resolution reconstruction algorithm of weather radar data based on adaptive regularized sparse representation (ARSR) is proposed, which incorporate autoregressive (AR) and non-local (NL) adaptive regularization terms in the sparse representation framework by effectively using local and non-local information from weather radar echo data to improve the reconstruction effect of the edge and detail information of the radar echo. Experimental results show that the ARSR algorithm substantially outperforms common interpolation for ×2 and ×4 resolution improvement in terms of both objective evaluation metrics and visual quality, especially for the reconstruction of the edge and detail information of the weather radar echo.