Real-time monitoring of traffic parameters

Abstract

This study deals with the problem of rea-time obtaining quality data on the road traffic parameters based on the surveillance camera data. The purpose of the paper is to develop a system to collect data on the traffic flow structure and to determine the traffic flow speed and direction in real-time. Our solution is based on the use of the YOLOv3 neural network architecture and open-source tracker SORT. To increase the accuracy of detection and classification, we used multi-scale prediction with an increased number of anchors. To determine the speed, we used a matrix of the perspective transformation of the source image to geographical coordinates. To train the neural network, we marked over 6,000 images and performed augmentation, which allowed us to increase the dataset to 60,000 images. Checking the system at night and in the day showed an absolute percentage accuracy of counting vehicles of no less than 92%. The error in determining the vehicle speed by the projection method, taking into account the camera calibration, did not exceed 2.74 m/s. The presented study allows us to generate big data for the intelligent transport systems decision-making system in real-time and to lower the requirements for peripheral equipment.

Improving road safety in small towns based on a systematic approach

“…Due to the intellectualization and digitalization of all processes and systems, the intelligent transport system (ITS) is one of the main tools of improving road safety. The introduction of ITS is associated with the use of modern communication facilities, the development of data processing algorithms, monitoring and control methods, using onboard computers, sensors and other technical means, which are equipped with modern vehicles and infrastructure facilities [5,6]. Effective management decisions are formed on the basis of information received in real time [7,8].…”

Predicting and Managing Lane Capacity at a Signal-Controlled Intersection

“…Другим примером компромисса между скоростью и точностью является архитектура YOLO [4,22], которая специализирована для обнаружения транспортных средств с помощью кластеризации якорей, дополнительной нормализации потерь и стратегии слияния многоуровневых функций. Основываясь на задаче обнаружения объектов в реальном времени, нейронная сеть YOLO v4 способна обрабатывать максимальное количество кадров в секунду, при этом она не сильно теряет в точности.…”

“…Разработанная система мониторинга была протестирована на нескольких перекрестках в г. Челябинске, показав среднюю ошибку подсчета транспортных средств 5,5 %. Погрешность определения скорости автомобиля проекционным методом с учетом калибровки камеры на тестируемом перекрестке не превышала 1,5 м/с [4,26]. Для обнаружения и классификации ТС использована архитектура нейронной сети YOLOv4 и трекера с открытым исходным кодом (SORT).…”

“…Бурное развитие технологий в сфере распознавания объектов в реальном режиме времени нашло применение и в области регулирования трафика уличного движения [1,2]. С помощью методов обработки изображений и машинного обучения в режиме реального времени исследователи [3][4][5] предлагают эффективный способ оценки плотности трафика на перекрестке.…”

Using Deep Learning Technology to Realize the Automatic Control Program of Robot Arm Based on Hand Gesture Recognition

“…The technique "you only look once (YOLO)" [6][7] is used to perform real-time object detection with artificial intelligence technology. It can be applied to specific object detection [8][9], traffic flow calculation [10][11], self-driving applications [12][13], product defect detection [14][15], medical image analysis [16], etc. Creating a one-stage detection and classification process is the main developmental purpose of YOLO, which is different from most traditional detection networks.…”