Optimizing grouper catch efficiency using AI-controlled light fishing in Pamekasan

Abstract

Light fishing techniques are widely used to improve marine fish catches by adjusting the spectrum and intensity of light-emitting diodes (LEDs) to attract specific species. Grouper (Epinephelus spp.) are known to respond well to red light, although the optimal light intensity remains unclear. This study proposes a dual system consisting of a pulse-width modulation (PWM)-based red LED brightness controller and an artificial intelligence (AI) fish detection module using the YOLOv4 algorithm. LED brightness was varied at duty cycles of 50, 100, 150, 200, and 250 (1 kHz), producing luminous flux values between 120 and 650 lm. Field experiments conducted at the Pamekasan coast using 10 grouper samples showed that 80% of fish aggregated at a duty cycle of 150 (~420 lm at 1.5 m). ANOVA results (p<0.05) indicated significant differences among light intensities, with the highest attraction observed at this level. These findings suggest the existence of an optimal light intensity threshold. Further studies with larger samples and multi-site experiments are required to validate these results and evaluate ecological impacts.