This innovation is a smart automated fish-feeding system specifically developed for small-scale tilapia farmers. It integrates a computer vision model (YOLOv8, trained using Roboflow) to estimate fish biomass in real-time via an ESP32-CAM and an IoT-based water quality monitoring system, utilizing sensors for dissolved oxygen, temperature, pH, ammonia, and salinity, all connected to an Arduino Uno R4 Wi-Fi microcontroller. The visual and environmental data are used to dynamically control feeding based on actual fish biomass. Notable features include a feeder hopper made from recycled materials, a user-friendly design (casing sensor), and real-time dashboard monitoring through the custom-built AquaConnect interface. This innovation has been successfully implemented at 2 tilapia farms (Kuala Pilah and Chuah), where it has demonstrated significant positive impacts over 1 month of operation:

• Feed Cost Reduction: Up to 20% reduction in feed waste, improving Feed Conversion Ratios (FCR) from 1.85 to 1.55 (Chuah) and from 1.7 to 1.42 (Kuala Pilah).
• Labor Savings: Reduces manual feeding labor by up to 70%, allowing farmers to focus on water quality management and post-harvest operations.
• Improved Fish Growth: Achieved an increase in daily fish growth rate by 4.9% to 8.3%.
• Environmental Sustainability: Minimizes water pollution by reducing excess organic waste and ammonia through precise feeding and continuous water monitoring.
• Scalability and Accessibility: The system is adaptable to different pond setups and affordable for small-scale farmers due to its use of low-cost and recycled components.