Fishency has developed Fishency360, a hardware and software solution based on computer vision and machine learning to capture 360-degree views of individual fish at different depths in the cage. High-resolution images of individual fish are used to detect and categorize sea lice and evaluate the quality of a sample (fish maturity, K-factor). Moreover, the images are used to measure growth and to detect early stages of fish health issues such as winter ulcers, cataracts, scale loss, and other skin damages. Fishency360 can be applied to traditional open sea farming as well as land-based or offshore operations. Fishency has so far had two full-scale prototype tests in open sea cages to prove our approach and confirm that a large fish sample swims passively through our system every day. We have since our last trial refined our hardware, and completed the first version of our software to enable more advanced tests and pilots. The current Technology Readiness Level (TRL) is assessed to be at 6, while we also find arguments in support of TRL 7. At present, we have 3 signed LoIs and 2 signed trial agreements that will be put into action as part of the EIC Accelerator program. Fishency needs access to a large amount of sea lice data in addition to data related to other fish welfare indicators to continue developing our machine learning algorithms, increase accuracy and validate our quality of service. Fishency will improve its support to farmers by not only providing unique data, but also assist in the way the data is used to optimize production and improve fish welfare. Fishency360 will have to be validated also by regulators to replace manual counting fully.
Project leader: Flavie Gohin
Institution: FISHENCY INNOVATION AS