Use of One-Stage Detector and Feature Detector in Infrared Video on Transport Infrastructure and Tunnels

Abstract

This article presents the use of the combination of the object detection method and feature detector in an infrared video on traffic infrastructure and in tunnels. The theme of the paper is the validation of vehicle detection and its classification using infrared video streams. In addition, the article focuses on the use of a feature detector and object detection to distinguish between vehicles with electric and combustion motors. The method suggests the use of a low-resolution thermal camera as an inexpensive extension of installed thermal camera technologies. The developed system has been verified for the applicability of vehicle detection and classification using object detection methods and their application in transport infrastructure and tunnels. It also presents a method for distinguishing propulsion units into electric and internal combustion; both systems’ conclusions are then statistically verified. The application of the system is evident in regional traffic management systems, including safety applications for traffic control in tunnels. Categorizing vehicles provides valuable information for higher levels of traffic management, toll systems, and municipal database systems, as well as for a preventive system for estimating vehicle conditions and their potential of fire in tunnels.