Real-Time Panoramic Multi-Target Detection Based on Mobile Machine Vision and Deep Learning

Abstract

Research on computer vision technology based on deep learning has become an important research direction in the field of artificial intelligence. Among them, the target detection task of images and videos has occupied a pivotal position in many intelligent vision research and applications. The purpose of this article is to learn from the existing object recognition and target detection technology foundation, to conduct in-depth research on the current traffic patrol deficiencies, to achieve real-time patrol and monitoring of key roads. This article chooses to use the target detection algorithm of Faster R-CNN in deep learning, applies drones and mobile machine vision to traffic patrol inspection, and detects various targets in the PASCAL VOC data set. The AP of the bus is 68.20%; The AP of the car is 75.50%; The moving object recognition and tracking method based on Faster R-CNN and Hungarian matching method is verified, and traffic flow data can be calculated to realize real-time monitoring of road traffic. The research in this paper helps to overcome the defects of human eye recognition in the traditional monitoring system and improve the efficiency of traffic inspection.