Abstract:

The results presented are from the NOAA EPP CSC NERTO graduate internship project conducted under the mentorship of Tim Battista, National Centers for Coastal Ocean Science (Silver Spring). This NERTO experience aligns with the NOAA Cooperative Science Center in CESSRST-II, supporting the Center’s goal of conserving and restoring coastal and marine ecosystems. This project, titled Quantifying Seafloor Visibility in Underwater Images: A Deep Learning Approach, addressed the research question: how can deep learning models be used to quantify turbidity in images of the seafloor. Images of the Gulf’s seafloor from field missions related to the Mesophotic and Deep Benthic Communities project were used to train a deep learning model to recognize different levels of turbidity. The model was then used to rank images based on the presence of turbidity to allow for a deeper understanding of what percentage of the data is usable for downstream analysis.

The work provides value to the scientific community and program stakeholders by providing an efficient method for data quality analysis in the case where multiple terabytes of data are being collected. Measuring data quality as it is being collected in the field allows the field team to monitor if the collected data is usable for downstream modelling. If not, real-time changes can be made so make efficient use of resources and improve the quality of collected data. Through this internship, the student deepened their understanding of NOAA mission areas and gained enhanced skills in data analysis, modeling, coding, and remote sensing.