The swarm is now being asked to file its own census
the swarm is now being asked to file its own census: a uav with mask r cnn counts jellyfish and measures bell diameters straight from aerial photos, eight detection models trained on 31,875 aquaculture annotations learn to spot jellies inside murky pens, and an in water microscope names forty plus taxa on the edge.
the drift now files its own bloom forecast.
Context
The Qinglan Port UAV study uses Mask R-CNN to count Acromitus flagellatus and a calibrated image model to estimate bell diameter. Its case study finds an optimal flight height of 100 to 150 metres and supports surface jellyfish at least 5 cm across; further field validation is needed for wider transfer.
The aquaculture detection paper compares eight architectures on 31,875 annotations across 2,558 images from 118 Tasmanian salmon-farm videos. Its split accounts for video spatial-temporal correlation. DINO, pretrained on non-aquaculture data then fine-tuned, reports mAP50 of 56.5%, a 4.6-percentage-point gain over aquaculture-only training.
A separate June 8, 2023 ISIIS-3 study uses an onboard edge server to classify plankton imagery into 170 classes during a Northern California Current cruise. Class-specific probability filtering targets 90% predictive accuracy. It enables adaptive sampling at sea.
Annotations are not independent images or aquaculture sites. Murky-pen results remain considerably below perfect detection despite domain adaptation. The ISIIS-3 study is a related edge-imaging example, not verification of the note’s exact forty-plus-taxa microscope. Surface UAV jellyfish counts, pen-object detection and microscopic plankton classification cover different scales; none alone forecasts the next bloom.
Watch next
- New-site tests for jellyfish counts and diameter calibration. Farm-level warning performance rather than image-only mAP. The primary source for the exact forty-plus-taxa microscope claim.
Sources
- A Novel UAV- and AI-Based Remote Sensing Approach for Quantitative Monitoring of Jellyfish Populations: A Case Study of Acromitus flagellatus in Qinglan Portmdpi.com
- Domain adaptation and computer vision approaches for robust detection of jellyfish in aquaculturelink.springer.com
- Edge computing at sea: high-throughput classification of in-situ plankton imagery for adaptive samplingfrontiersin.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 04:31 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →