AI/ML-Powered Recyclable Bottle Sorting System

This project showcases the integration of cutting-edge AI/ML techniques with industrial automation to solve a real-world environmental challenge. The goal was to develop a high-speed, high-accuracy sorting solution capable of identifying and classifying recyclable bottles plastic, metal, and glass in real-time under industrial conditions.

IHub-Data Support & Technology Enhancement

Through the support and collaboration with iHub-Data, the system was significantly optimized. Uniform illumination was introduced to ensure consistent imaging, eliminating lighting-based distortions. A hardware upgrade replaced rolling shutter cameras with high-speed global shutter cameras for improved motion capture. Additionally, advanced AI/ML detection models were integrated and customized for real-time object classification and sorting, enhancing performance across varying material types like plastic, glass, and metal.

Impact & Results

The optimized solution delivered a 2X speed boost, raising the sorting rate from 120 to 240 items per minute, meeting the client’s operational goals. The accuracy improved dramatically, increasing by 25% from 70% to 95%, enabling high-confidence sorting decisions. The system now supports industrial-grade reliability, making it robust, scalable, and ready for 24×7 continuous operation in real-world production environments.