Traffic Sign Classification Using Histogram of Oriented Gradients in Autonomous Driving Assistance

Traffic sign classification is a crucial component in autonomous driving assistance systems, enabling vehicles to interpret road signs for safe and intelligent navigation. Effective and timely recognition of traffic signs ensures improved driver support and road safety. However, existing deep learning-based methods often require large datasets, high computational resources. They can be sensitive to variations in lighting, occlusion, and scale, which limits their real-time applicability in embedded systems. To address these challenges, this paper proposes a framework based on Histogram of Oriented Gradients for feature extraction combined with a Support Vector Machine classifier (HOG-SVM). HOG efficiently captures the structural and edge features of traffic signs, while SVM offers robust classification with low computational complexity. This lightweight HOG-SVM approach is well-suited for real-time applications in Advanced Driver Assistance Systems (ADAS), especially in resource-constrained environments. Experimental results demonstrate that the proposed method achieves high accuracy in traffic sign classification with reduced processing time and enhanced robustness to noise, rotation, and illumination changes. The system is reliable, fast, and adaptable, making it practical for deployment in autonomous vehicles. The proposed method gradually improves the accuracy by 96%, image processing time (300-400 ms), consistent image resolution of 32×32, and feature vector size of 3780.

Перевод пока недоступен