Object Detection Based on Digital Twin-Based Road Scenarios
Zhang Yu, Chen Huimin, Wang Weihan · 2024
With the rapid development of artificial intelligence and digital technology, digital twins are becoming more and more widely used in various fields, and have high application value in industrial production, architectural design and other fields. There are few datasets of extreme scenarios used in traditional object detection and vehicle autonomous driving training, resulting in unsatisfactory results of object detection and autonomous driving technology in complex environments. In this paper, a digital twin model of the road scene is built based on the rules of City Engine and the Unreal Engine 4 platform, which can be used to build a model of complex road scenes and shorten the model construction cycle. Based on the constructed digital twin model, the target detection task is carried out, and it is verified that the target detection and autonomous driving tests can be carried out in the digital twin model. It provides ideas for solving the problem of insufficient testing of autonomous driving and object detection tasks in complex and extreme scenarios.