Real Time Vehicle Detection and Counting Using Deep Neural Network

The number of vehicles on the road is increasing rapidly. By using the number of vehicles as valuable data for detecting traffic congestion, it is required to detect vehicles on the road accurately and quickly, which is useful for traffic management. This paper presents a model to detect different vehicles and count them in a given video frame. With the help of YOLOv4 we can detect the different types of vehicles in the given video. Furthermore, it uses Deep SORT algorithm to help count the number of vehicles passing in the video effectively. Performance parameters like IOU and map are calculated for measuring the performance and validation of this work. After the implementation of algorithms, based on the results of YOLOv4 testing resulted in a detection accuracy rate with mAP of 84.50% where the combination of YOLOv4 with the Deep Sort algorithm can detect, track and count four types of vehicles.

Keywords: YOLOv4, Deep Sort, Detection, Tracking

Suggested Citation: Suggested Citation

Mittal, Dr. Himani and Chhabra, Prachi and Kedia, Harsh and Agarwal, Aahan, Real Time Vehicle Detection and Counting Using Deep Neural Network (August 5, 2022). Proceedings of the Advancement in Electronics & Communication Engineering 2022, Available at SSRN: https://ssrn.com/abstract=4186826 or http://dx.doi.org/10.2139/ssrn.4186826