Real-Time, Deep Learning Based Wrong Direction Detection

Saidasul Usmankhujaev,Kwon Jang Woo,Shokhrukh Baydadaev

doi:10.3390/app10072453

Abstract

In this paper, we develop a real-time intelligent transportation system (ITS) to detect vehicles traveling the wrong way on the road. The concept of this wrong-way system is to detect such vehicles as soon as they enter an area covered by a single closed-circuit television (CCTV) camera. After detection, the program alerts the monitoring center and triggers a warning signal to the drivers. The developed system is based on video imaging and covers three aspects: detection, tracking, and validation. To locate a car in a video frame, we use a deep learning method known as you only look once version 3 (YOLOv3). Therefore, we use a custom dataset for training to create a deep learning model. After estimating a car’s position, we implement linear quadratic estimation (also known as Kalman filtering) to track the detected vehicle during a certain period. Lastly, we apply an “entry-exit” algorithm to identify the car’s trajectory, achieving 91.98% accuracy in wrong-way driver detection.

Highlights

With the development of the fourth industrial revolution, the role of the intelligent transportation system (ITS) is becoming more and more crucial for ensuring the safety and efficiency of drivers
As soon as it reaches area C, image and information about the car were sent to the monitoring center using the File Transfer Protocol (FTP) and Transmission Control Protocol (TCP)
We propose a combined state-of-the-art you only look once version 3 (YOLOv3) method with an updated linear quadratic equation to detect and track vehicles

Summary

Introduction

With the development of the fourth industrial revolution, the role of the intelligent transportation system (ITS) is becoming more and more crucial for ensuring the safety and efficiency of drivers. CCTV footage certainly cannot be a perfect solution for all traffic violations. It is widely and increasingly used by government officials to identify serious violations, along with evidence, testimonies, and documents. Wrong direction estimation is not a contemporary problem, and we we describe describe and and implement implement different different techniques techniques in in computer computer vision vision to to compare compare different different detection detection and and tracking algorithms for vehicles. These techniques include well-known machine learning and deep tracking algorithms for vehicles.

Methods

Results

Conclusion