Bayesian Rain Field Reconstruction using Commercial Microwave Links and Diffusion Model Priors

2026-05-06

AI Summary

This study addresses the limited accuracy of existing commercial microwave link (CML)-based rainfall field reconstruction methods under non-uniform precipitation, which arises from oversimplifying path-integrated observations as point measurements. The authors formulate the reconstruction task as a Bayesian inverse problem and, for the first time, incorporate a pre-trained diffusion model as a spatial prior. This approach enables high-fidelity rainfall field recovery without additional training, leveraging plug-and-play optimization, sequential Monte Carlo, and replica exchange sampling for posterior inference. By better preserving the statistical characteristics of rainfall, the proposed method significantly outperforms current baselines on both synthetic and real-world datasets, with particularly notable improvements in reconstructing highly non-uniform precipitation patterns.

Abstract

Commercial Microwave Links (CMLs) offer dense spatial coverage for rainfall sensing but produce path-integrated measurements that make accurate ground-level reconstruction challenging. Existing methods typically oversimplify CMLs as point sensors and neglect line integration relating rainfall to signal attenuation, resulting in degraded performance under heterogeneous precipitation. In this work, we view rain field reconstruction as a Bayesian inverse problem with Diffusion Models (DMs) as high-fidelity spatial priors. We show that diffusion models better preserve key rainfall statistics compared to censored Gaussian processes. Framing rainfall estimation as a Bayesian inverse problem with a DM prior enables training-free posterior sampling using a broad family of methods, including Plug-and-Play, Sequential Monte Carlo, and Replica Exchange methods. Experiments on synthetic and real-world datasets demonstrate consistent improvements over established CML-based reconstruction baselines.

Problem

• rain field reconstruction
• Commercial Microwave Links
• Bayesian inverse problem
• spatial rainfall heterogeneity
• path-integrated measurements

Innovation

• Diffusion Models
• Bayesian Inverse Problem
• Commercial Microwave Links
• Rain Field Reconstruction
• Path-Integrated Measurements