Multi-instance vehicle dataset with annotations captured in outdoor diverse settings

Published: 7 March 2023| Version 2 | DOI: 10.17632/5d8k5bkb93.2


We collected and annotated a dataset containing 105,544 annotated vehicle instances from 24700 image frames within seven different videos, sourced online under creative commons license. The video frames are annotated using DarkLabel tool. In the interest of reusability and generalisation of the deep learning model, we consider the diversity within the collected dataset. This diversity includes changes of lighting amongst the video, as well as other factors such as weather conditions, angle of observation, varying speed of the moving vehicles, traffic flow, and road conditions etc. The videos collected obviously include stationary vehicles, to perform the validation of stopped vehicle detection method. It can be noticed that the road conditions (e.g., motorways, city, country roads), directions, data capture timings and camera views, vary in the dataset producing annotated dataset with diversity. the dataset may have several uses such as vehicle detection, vehicle identification, stopped vehicle detection on smart motorways and local roads (smart city applications) and many more.


Steps to reproduce

The dataset is acquired from online sources with creative common license. The major limitations with existing similar datasets is unavailability of annotations where vehicles are moving in both directions. Likewise, the diversity is limited in the existing datasets particularly, there is no dataset for time series vehicle tracking in diverse conditions. This dataset might be useful to finetune deep learning models such as YOLO, R-CNN etc., for vehicle detection as well as real-time tracking tasks.


Liverpool John Moores University


Tracking Algorithm, Vehicle, Smart City, Transport Highway, Deep Learning, Object Tracking (Computer Vision)