This post explores the use of computer vision in monitoring traffic levels in Norway. By fine-tuning pre-trained CNN and ViT models on public traffic cameras, the author aims to automate the process of estimating traffic levels well before the API updates. The ViT model performs better compared to other models, and a dataset of traffic camera images is used for training.
Table of contents
Working with Computer Vision: Monitoring Traffic in NorwayIntroductionTraining the ModelSort: