Flow-through optical imaging for quantification and classification of organic and inorganic particulates in aquatic environments

Aquatic environments host both natural and alien particulates ranging from essential ecosystem contributors such as zooplankton and larval fish to pollutants such as microplastics. Across the range of organic and inorganic particulates, a common feature is a high degree of spatial and temporal variability. Understanding that variability at the fine scale is essential to effective management. However, available sampling methods, and particularly legacy net-based methods, are limited in their ability to operate across the spectrum of applications and/or to deliver the required resolution. We describe a flow-through optical imaging system capable of processing relatively large volumes (up to 23 liters per minute) of water, that is easy-to-use, flexible for deployment in the laboratory or in field settings ranging from small streams to the open ocean, and that classifies and quantifies particulates ranging in size from ~1-50 mm in near-real-time. Both grayscale and color models are available. The grayscale version works well in marine environments where target organisms tend to be translucent. The color option is ideal for imaging microplastics, larval fish, and freshwater benthos where surface features and colors contribute importantly to classification success. Validation testing supports classification accuracy >80% when backed by a sufficiently replete training library.

Presentation Preference: Poster

Primary Presenter: William Arnold, OceanSpace LLC (oceanspacesensors@gmail.com)

Authors:

William Arnold, OceanSpace LLc  (oceanspacesensors@gmail.com)

Kurt Kramer, OceanSpace LLC  (kkramer.oceanspacesensors@gmail.com)

Milton Dean, OceanSpace LLC  (mdeanresearch@gmail.com)

Eric Steimle, OceanSpace LLC  (esteimle@gmail.com)

Karsen Henwood, University of South Florida College of Marine Science  (KarsenMM@gmail.com)