Northern Gulf of Alaska: DPI-3
Field proven in the Northern Gulf of Alaska
Pelagia grew from direct experience operating DPI-3 and processing imagery from its three ISIIS-DPI plankton cameras for the University of Alaska Fairbanks and the Northern Gulf of Alaska Long-Term Ecological Research program.
ISIIS-DPI processing at Icy Seas
Hannah Kepner, Digital System Specialist
Hannah focuses on processing imagery from the ISIIS-DPI plankton cameras carried by DPI-3, bringing her experience with in situ plankton imagery, machine-learning classification, and Northern Gulf of Alaska gelatinous zooplankton research to Pelagia's scientific workflows.
Explore Hannah's ISIIS-DPI work →A deployment vehicle carrying a complete observing system.
DPI-3—often shortened to “DPI” and nicknamed “Big Friend”—is a Bellamare-built, actively controlled undulating towed vehicle used by the Northern Gulf of Alaska LTER program. It carries three ISIIS-DPI plankton cameras together with acoustic and environmental instruments. The cameras produce high-resolution line-scan shadowgraph imagery, while DPI-3 provides the controlled deployment platform and carries the broader observing payload.
Organisms and particles are observed in place, without collection or physical handling, as DPI-3 crosses depth layers and hydrographic gradients.
Each image is part of a spatially and temporally indexed observation stream. Time, position, vehicle depth, camera, sampling context, environmental measurements, and processing history all matter alongside the pixels.
A data stream measured in millions of liters and hundreds of terabytes.
The reported UAF/NGA operating configuration creates an unusually demanding image-processing problem. Exact rates vary with camera and acquisition settings, but the scale is constant: sustained shipboard acquisition, sparse targets within an immense background, and far more imagery than a science team can inspect manually.
Instrument to data
One traceable operational workflow.
Pelagia reflects the full lifecycle: reliable acquisition, high-throughput storage and transfer, distributed processing, automated detection and classification, scientific review, and analysis-ready export. Every derived region, mask, measurement, and classification can remain traceable to its source imagery, acquisition metadata, processing configuration, and review state.
Fine-scale imagery changes what scientists can see.
Traditional plankton nets remain essential, but they can miss fine spatial structure and damage fragile gelatinous organisms. Imagery from the ISIIS-DPI plankton cameras preserves organisms in their oceanographic setting. Research from the 2022–2023 DPI-3 deployments found aggregations of ctenophores, hydromedusae, and siphonophores around frontal features, reported previously undetected species records, and showed that gelatinous zooplankton were more prominent in the ecosystem than historical sampling had indicated.
Biological distributions
Map abundance, size, and fragile gelatinous zooplankton across depth and distance.
Environmental context
Relate biological patches to temperature, salinity, oxygen, acoustics, and frontal structure.
Defensible records
Create quality-controlled, georeferenced observations for ecological analysis and model training.
Experience across the lifecycle
Software designed from the instrument outward.
Pelagia's lead developer served as a chief DPI-3 operator on NGA cruises and led image-processing and technical-computing work for its ISIIS-DPI plankton cameras—from high-throughput shipboard acquisition through segmentation, machine-learning classification, scientific review, and analysis-ready export.
Responsibilities
- Instrument operations and at-sea deployment support
- Raw-data acquisition, storage, and transfer architecture
- Segmentation and machine-learning classification
- Processing pipeline and compute infrastructure
- Quality control and analysis-ready records
- Technical support for students and collaborators
Explore the science and the program.
Read the public research behind the NGA observations or learn more about the long-term study of this productive subarctic ecosystem.