Environmental Informatics and Neutron Activation Analysis for Addressing Heavy Metal Contamination Concerns at the Community/Individual Scales.
Synopsis of project aims:
Develop a Machine Learning (ML) method for automated gamma-ray spectral identification and uncertainty quantification of metal concentrations in NAA. We will apply the convolutional neural network-based method recently developed for gamma spectroscopy, and also explore other methods.
Establish sampling, tracking and data management workflow. We will use a platform to generate the International Geo Sample Number (IGSN) identifier as well as create a database to store NAA data, and associated metadata (e.g., sampling time/locations, sample picture). All the datasets will be in a machine-readable format. In addition, we will explore how to address the privacy issues of such environmental datasets (e.g., anonymizing, rounding coordinates).
Develop ML toolsets to identify and visualize the spatial patterns of metal concentrations at the regional/national scale, and to attribute their heterogeneity with environmental and anthropogenic factors. It leverages the existing environmental informatics framework that Professor Wainwright has developed.
