At present, the main difficulties in the hatching and growing process of fry are that the size of fry is extremely small and difficult to analyze, various complex water quality data need to be judged, and the production process of algae and live-feeding organisms is complicated. As a result, fry production is still heavily reliant on manual labor. For example, tubular rotifer filtration can efficiently produce live-feeding organisms, and use underwater visual recognition to analyze the size of fry, but the problem of algae recognition and cultivation has yet to be broken through. The main technical barrier for algae at the moment is density detection. Although the AI visual recognition system has an error rate of 3-9%, manual recognition has a range of 12-25%. However, mixed microalgae (Picochlorum sp. S1b) culture density can exceed 109 ind/mL, but it cannot be identified if it exceeds 105 ind/mL, and the algae sample volume is only 20cc, and the relevant dilution equipment is mostly manual and cannot be automated. The capacity of the equipment with automation is too large. As a result, this study employs solenoid valves, peristaltic pumps, and an automatic dilution device with a programmable logic controller (PLC) control system that can achieve automatic dilution 100 times with a 4.3% error rate (Figure 1). The system allows the intelligent fry production system to make accurate decisions, consistently automate algae production, and collaborate with other systems to achieve the goal of automated fish production.
Keywords: automation, programmable logic controller, microalgae