Artificial intelligence garbage sorting vehicle based on deepstream

Abstract

In this paper, an intelligent garbage sorting vehicle based on the upper computer jetson nano and STM32 is designed. The upper computer jetson nano collects images of garbage objects through a usb camera, and the collected data is applied to the deep learning algorithm YOLO model to complete the garbage identification and classification function, and the radar is used to locate and avoid obstacles. Through the communication between the upper computer and the lower computer STM32, the trolley is driven to the specified position, and the garbage is sorted to the corresponding trash can by the mechanical arm. STM32 with the Internet of things function, can display the quantity and type of each garbage on the mobile app, so that the administrator can timely operation and maintenance. In this paper, the hardware and software design and joint debugging of the above system are completed from the reality.