<![CDATA[The increasing volume of waste in Indonesia is a major environmental challenge due to poor source segregation. This study develops a Smart Garbage system based on IoT and computer vision to classify waste into metal, organic, and inorganic categories with real-time monitoring. The system integrates an ESP32 microcontroller, inductive proximity sensor, webcam, Teachable Machine, Python, and the Blynk platform. The sensor detects metal waste while image classification distinguishes organic and inorganic waste. Results are sent to ESP32 via serial communication to control the sorting mechanism. Testing evaluated detection accuracy, confidence threshold, lighting conditions, and communication delay. The inductive sensor achieved 100% accuracy for metal detection. The optimal confidence threshold was 0.8 with 90% classification accuracy. The average confidence score reached 0.89 under good lighting and latency averaged 4.26 seconds. The Blynk platform enabled real-time monitoring. Overall IoT and computer vision improve waste sorting efficiency and system performance effectiveness overall improved.]]>
Copyrights © 2026
