AI Insight
This study introduces a Microcanonical Langevin Monte Carlo (MCLMC) sampler for reconstructing radiation images from radiological mapping data while also quantifying uncertainty in the results. Unlike traditional methods that only provide point estimates, MCLMC generates probabilistic distributions of radioactivity levels and converges to accurate posterior distributions in approximately 10 seconds for images containing 1,000-10,000 pixels when run on GPU hardware. The method was validated on both synthetic data and real-world distributed source measurements, showing improved accuracy compared to Maximum Likelihood Expectation-Maximization approaches with reduced risk of over- or under-fitting.
Why it matters
This advancement enables emergency responders and environmental managers to quickly obtain not only radiation distribution maps but also confidence intervals for those estimates during nuclear incidents or contamination events. The combination of speed and uncertainty quantification can significantly improve decision-making in time-critical radiological emergencies where understanding the reliability of measurements is as important as the measurements themselves.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract: Radiological mapping plays a critical role in nuclear emergency response and environmental management activities. A radiation image, representing the spatial and intensity distribution of the radioactivity, is reconstructed from the radiation data and the associated contextual information. Typical image reconstruction methods, such as Maximum Likelihood Expectation-Maximization (ML-EM), only provide point estimates of the pixel or voxel activities without associated uncertainties. Here, we apply a new Microcanonical Langevin Monte Carlo (MCLMC) sampler for radiation image reconstruction and uncertainty quantification. The MCLMC sampler properties are first tested with synthetic radiation images. Methods to obtain the radiation distribution estimate and the associated uncertainty from the samples drawn by MCLMC are discussed. Given sufficient measurement statistics, the radiation distribution estimate obtained from MCLMC results closely resembles the ground truth with less risk of over- or under-fitting compared to ML-EM. When MCLMC is run in parallel on a GPU, the samples can converge to the posterior distribution in about 10 seconds for an image with $10^3$–$10^4$ pixels, which is significantly faster than other comparable Markov Chain Monte Carlo (MCMC) samplers. We also tested MCLMC on a dataset from a real distributed source radiological mapping campaign, and the reconstructed results agree well with the expected activity map. The fast MCLMC sampler therefore enables improved imaging accuracy and prompt uncertainty quantification for reconstructed radiation images, which can better inform decision-making in response to radiological events.