2 Fish welfare and health monitoring
2.1 Behaviour monitoring and classification
The task is aiming the detection of abnormal behaviors using Computer Vision and Artificial Intelligence. Specifically, the study of the underwater videos and the analysis of the motion allows an association of the fish school reaction with dynamic changes of their environment. These changes may occur due to weather conditions variability, changes in feeding practices, physiological status, etc.
As a first step, the tracking of an individual from a fish group is performed as long it appears in a minimum set of subsequent frames. The instant velocity of each fish is estimated, aiming for an estimation of the average velocity for the whole group. The group velocity is then correlated to the monitoring conditions (feeding, weather, etc). The final objective is to detect abnormal changes in the movement pattern of the fish group.
Sample results of fish detection, motion tracking & velocity estimation:
If you want to test the application, contact us and we shall provide you with further details
Demo Video of the Application
2.2 Health monitoring
The aim of the current work is the development of a system that will detect disease precursors using hyperspectral imaging sensors. The expected outcome is the identification and classification of different epidermal alterations using their hyperspectral signature that can be used as a warning sign and facilitate acute intervention before the disease is widely spread.
Preliminary trials with controlled and manipulated fish (undergone, netting, air-exposure, mucus and/or scale removal) are used to determine a methodological approach.
First data shown variations in intensity in the range of 700 to 1000 nm wavelength, at areas with a skin damage
Deliverables
| Behavioral estimation | Health evaluation |
| D1.1.1 End-user requirements and technical specifications of the sub-systems | |
| D1.2.1 Content and specifications of the databases and experimental data reference tables | |
| D3.1.1 System for optical fish behavior detection (hardware and software) | D3.2.1 System for fish health evaluattion using hyperspectral cameras |
| D3.1.2 Operation manual of the optical fish behavior detection system | D3.2.2 Operation manual of the fish health evaluation system |
| D4.1.1 Trial and system evaluation results | |
| D4.1.2 Technical, financial and utilization study of the systems | |
| D4.2.1 Report on the results’ dissemination actions | |
