Design of multi-garbage classification system based on edge AI

A multi-garbage classification system based on edge artificial intelligence (AI) has been designed for improving the real-time capability and accuracy of embedded garbage classification intelligent terminals. The system firstly employs the industrial assembly line methodology to facilitate the physical segregation of multi-garbage through the utilization of double conveyor belts.Secondly, a high-performance edge AI platform is employed for the acquisition and processing of images, and a network architecture based on the MobileNetV2 lightweight feature extraction network and the SSD (Single Shot MultiBox Detector) target detection algorithm is constructed to accurately identify garbage.Finally, the multi-steering engine linkage control is designed for accurate garbage classification. The structural design of the system and the related hardware and software design are realized, After extensive testing, the experimental results show that the system can complete the identification and classification of 10 pieces of garbage, which belong to four categories, within 40 seconds. with an average accuracy of over 98%.The system integrates control technology, edge AI technology and sensor technology, with high operational stability, fast response speed and high recognition accuracy, providing an effective solution for waste sorting in public places.