In some countries, semi-trucks are not allowed to use the inner lane on motorways during daytime. This project detects this situation and saves the evidence in a PostgreSQL database so the public authorities can initiate proceedings against the offender. In this project, I've built and trained a deep neural convolutional network to classify vehicles using PyTorch. OpenCV is used to process and filter frames and to localise moving objects on video.
CNN model is trained on the dataset of IEEE BigData Cup Challenge 2022. The images are prepared for object detection problems so I've cut every object to make it into an object classification dataset. The project uses the car_front and truck_front images of the dataset.