Development and performance evaluation of an automatic size-sorting system for catfish seeds using photodiode sensors

Abstract

In catfish farming, uniform seed size is crucial for ensuring balanced growth and minimizing competition for feed. Generally, size sorting is performed manually through visual observation and net separation, which is labor-intensive, time-consuming, and often causes stress or injury to fish. To address these limitations, this study aimed to develop and evaluate a real-time, low-cost automatic sorting system for live catfish seeds. The proposed system utilizes photodiode sensors and an Arduino-based microcontroller to detect variations in fish body length by interrupting a laser beam. Four photodiodes were arranged at specific distances to classify fish seeds into four size categories (<7 cm, 7–8 cm, 9–10 cm, and 11–12 cm). After classification, the system automatically directed each seed into the corresponding container. The results showed that the prototype successfully classified and sorted catfish seeds with an overall accuracy of 67.5%. In contrast, tests with PVC pipes under controlled conditions achieved 100% accuracy. These findings highlight the novelty of integrating size detection and direct sorting for live fish seeds, a feature not previously reported in the literature. Beyond its current limitations, this system provides a methodological framework for sensor-based aquaculture automation, offering potential for further improvements in accuracy, robustness, and application to other aquaculture species.