A Vehicle Detection Technique Using Binary Images for Heterogeneous and Lane-Less Traffic

Abstract

Nowadays, providing a low-cost traffic management system in developing countries or in heterogeneous and lane-less traffic conditions is highly essential. It can help to manage traffic congestion, save fuel, save travel time, and enhance user safety. By keeping these as an objective, this article presents a vehicle detection method in heterogeneous and lane-less traffic by extracting a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">binary image</i> from a discrete sensor array. The <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">binary image</i> is formed with a logic 1 or 0, which are recorded based on the occupancy status of the vehicles in an observed zone. The proposed method is demonstrated with virtual loops in video and with an array of micro-LiDARs. The width and length of the vehicle are obtained from the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">binary image</i> , which is extracted from virtual loops in a video recording and classified the vehicles. Similarly, the width and height information is obtained using an array of micro-LiDARs and classified the vehicles. The proposed method can easily be implemented with minimal storage, minimum cost, less bandwidth, and less computation complexity than the conventional methods, such as image processing or video processing-based vehicle classification. The proposed classification methods are mathematically derived, implemented, and measured performance over real traffic scenarios. It can be adopted automatically to high or light traffic scenarios by adjusting the distance between observation zones. The detection accuracy of 98% is observed while extracting data from video and 91.3% while using micro-LiDARs. The proposed works are compared with existing techniques.

Topics & Concepts