Michael Bollinger | Marine Scientist

My Role Autonomous Underwater Vehicles My work as part NOAA Mesophotic and Deep Benthic Communities (MDBC) was apart of the Mapping, Ground truthing and Modeling (MGM) team. As a part of this team of this team, we employed REMUS AUV's with both Kraken MINSAS 120 Synthetic Aperture Sonar (SAS), and VOYIS RECON LS Imaging and Laser Line Scanner (LLS) payloads. I've worked alongside engineers at Woods Hole Oceanographic Institution, inventors of the REMUS AUV, to successfully collect these data and learn the ins and outs of AUV operations. Imagery Analysis Along side AUV operations, I serve as a lead data analyst for the imagery that comes off the VOYIS sensor. I coordinate with AUV pilots to design missions to collect quality data, do the data management for the payload that collects terabytes of data on a single dive and develop tools to analyze and process the imagery. The team collects two main mission types, where the first is a photogrammetric development of an interesting area and the second is a "travelling salesman" where the AUV hits a number of randomly generated points for habitat model generation.  ​ Photogrammetry/Structure from Motion I take in the photogrammetric datasets and process them to correct the navigation. I then use Agisoft Metashape to create 3D reconstructions of the seabed. These reconstructions can turn into micro bathymetric analyses, biological cover censuses and educational outreach tools. ​ Substrate and Biologic Imagery Annotation My work in the annotation realm pairs my skills in imagery analysis with my degree pathway of Biology. I work with developers from CVisionAI to fine tune an annotation engine, TATOR. I identify all types of mesophotic biological cover and substrate types to feed into predictive habitat models.