title = “[Award] The Second Prize in National Computer Design Contest Tianjin Division” summary = " The China University Computer Design Contest (hereinafter referred to as the “Contest”) is organized by the Ministry of Education’s Higher School Computer Professional Teaching Guidance Committee, the Ministry of Education’s Higher School Software Engineering Teaching Guidance Committee, the Ministry of Education’s Higher School Computer Course Teaching Guidance Committee, and the Ministry of Education The school’s liberal arts computer basic teaching guidance subcommittee and China Education Television jointly sponsored. "

date = “2020-07-28T00:00:00+00:00” tags = [“Award”, “Deep Learning”, “5G”] authors = [“zhixianzhang”,“fanjiawen”]

[header] image = "" caption = ""

+++

This project is based on visual AI and 5G edge computing technology, through the functional customization of embedded devices, to realize the UAV intelligent patrol system. Through OpenCV for image processing, Yolo V4 for target detection and TensorFlow for deep learning, and through neural network algorithms, drones can autonomously extract features of motor vehicles and non-motor vehicles that violate traffic laws, and use embedded platforms and 5G The network transmission module empowers the UAV to automatically cruise and transmit to the cloud platform, and visualize the patrol mission through a customized App. It can replace the patrol officers to perform cumbersome patrol tasks that consume less time and effort, especially to solve the problem of indiscreet, large number of non-motor vehicles, and difficulty in checking and correcting the “one helmet and one area”. It can effectively and intelligently identify most traffic violations.