Toward Robust License Plate Recognition in Real-World Surveillance: A Systematic Literature Review and Conceptual Framework for Super-Resolution, Transformer-Based OCR, and Image Restoration

Abstract

Automatic License Plate Recognition (ALPR) has become a key technology for intelligent transportation systems, traffic surveillance, law enforcement, and smart-city applications. However, recognizing license plates captured in real-world surveillance remains challenging due to low-resolution imagery, motion blur, adverse illumination, compression artifacts, and other complex degradations that substantially reduce recognition accuracy. Recent advances in image super-resolution, image restoration, and Transformer-based optical character recognition (OCR) have shown promising improvements, yet these techniques are commonly investigated as independent components and evaluated using heterogeneous datasets and image-quality metrics. This paper presents a systematic literature review of 21 primary studies published between 2021 and 2025 following the PRISMA 2020 guidelines. The review identifies four research themes, synthesizes current challenges, and proposes a degradation-aware OCR-guided conceptual framework for future ALPR systems. The review provides a structured research roadmap for developing more robust ALPR solutions capable of operating under diverse real-world surveillance conditions.