Innovative aquaculture biometrics analysis: harnessing IR lasers and ToF cameras for microscopic fish larvae tracking

Abstract

Within the scope of aquaculture farm operation and research, monitoring fish larvae offers pivotal data about the operational conditions of the farm. For example, hypoxia may induce abnormal movements. Currently, precise monitoring of these diminutive entities (1 mm in size) hinges on superior water clarity and specialized equipment. While green laser may be preferred for extended range underwater imaging, it is visible to the fish. Hence it will disturb fish and potentially damage their vision system. This is of particular concern at our facility at the Harbor Branch Oceanographic Institute (HBOI). To address these challenges, our research has adapted a Time-of-Flight (ToF) camera, equipped initially with a 50mm lens, into a microscopic imager using an IR laser. This setup was capable of detailed but narrow depth field imaging, suitable for clear water conditions. Recent advancements have included transitioning to a 25mm lens, enhancing the camera’s ability to capture wider images (approximately 20 pixels wide for fish eggs) and observe finer details in medium turbidity conditions, though with a reduced depth field of 5mm. This modification has shifted the camera’s utility towards observing very small living organisms (100-200 microns) and reduced its effectiveness in depth measurement in highly turbid waters. This adaptation ensures more precise tracking of fish larvae and offers a fish-eye-safe imaging process due to the use of IR light. The integration of machine learning techniques further refines the system’s ability to accurately identify fish larvae in varying water conditions. Our approach presents a balanced solution, combining affordability, improved accuracy, and mindful consideration of the fish’s welfare, contributing positively to the field of fish larvae tracking.