A Concise Overview of Vehicle Detection Techniques

Abstract

Vehicle detection, a specialized subset of object detection, has gained significant importance in recent years, particularly in the realms of autonomous and assisted driving technologies. This field, while promising, grapples with several challenges including occlusion, scalability issues, and the complexity of real-world backgrounds. This paper sets out to provide a summary of recent state-of-the-art advancements in vehicle detection technology. First, it organizes vehicle detection approaches into three primary categories: classical methods, deep learning techniques, and hybrid approaches that combine elements of both. Within the deep learning category, the paper further distinguishes three subcategories: anchor-based methods, anchor-free methods, and attention-based techniques. Each of these approaches offers unique advantages and addresses different aspects of the vehicle detection challenge. Secondly, it provides a literature review of different papers on vehicle detection.