Deep Learning-Based Real-Time Traffic Sign Recognition System for Urban Environments

Abstract

A traffic sign recognition system is crucial for safely operating an autonomous driving car and efficiently managing road facilities. Recent studies on traffic sign recognition tasks show significant advances in terms of accuracy on several benchmarks. However, they lack performance evaluation in driving cars in diverse road environments. In this study, we develop a traffic sign recognition framework for a vehicle to evaluate and compare deep learning-based object detection and tracking models for practical validation. We collect a large-scale highway image set using a camera-installed vehicle for training models, and evaluate the model inference during a test drive in terms of accuracy and processing time. In addition, we propose a novel categorization method for urban road scenes with possible scenarios. The experimental results show that the YOLOv5 detector and strongSORT tracking model result in better performance than other models in terms of accuracy and processing time. Furthermore, we provide an extensive discussion on possible obstacles in traffic sign recognition tasks to facilitate future research through numerous experiments for each road condition.

Cox Regressive Outlier Robust Incremental Extreme Learning Machine based IoT aware Smart Car Parking and Route Optimization for Intelligent Transportation

“…Custom CNN model was used in [15] to categorize traffic signs with higher accuracy. An approach for traffic sign recognition was developed in [16] specifically designed for complex urban road environments. The method demonstrates higher accuracy and minimal latency.…”

Real-Time Traffic Light Recognition with Lightweight State Recognition and Ratio-Preserving Zero Padding

“…In particular, the TLD module is constructed following YOLOv5. The recent work in [15] also uses YOLOv5 to detect traffic signs that are different from traffic lights. Although it is a study to find different objects, it shows that YOLOv5 outperforms YOLOv3 and v4 when detecting objects in a driving environment.…”

“…Several works have been proposed for this purpose, classified into one-stage systems [7][8][9][10] and two-stage systems [11][12][13][14][15]. The one-stage system detects the location and state of traffic lights by using a single network.…”

“…The works in [12,14] adopted a histogram of oriented gradients (HOG) for the detection of traffic lights and also used a CNN for checking the status of traffic signals. In [15], the proposed system used deep learning networks for all components of the system to detect traffic signs.…”

A Survey on Cyber-Physical Security of Autonomous Vehicles Using a Context Awareness Method

“…In light of these vulnerabilities, the field of autonomous driving research has heavily emphasized refining these algorithms. Deceptive adversarial attacks, where small perturbations can mislead deep learning models, are of particular concern in the context of TSR and OC [19]. For instance, an attacker might subtly modify a TS's appearance, causing a TSR system to misinterpret it.…”