Pedestrian Detection Enhancement via Thermal-Visual Fusion
Thermal-Visual Fusion, Multispectral Pedestrian Detection, Infrared Imaging, Feature Fusion, Low-Light Perception, Intelligent Transportation
: Pedestrian detection remains a safety-critical perception task in autonomous driving, intelligent surveillance, and smart transportation, yet conventional visible-spectrum detectors still degrade sharply under low illumination, headlight glare, rain, haze, and partial occlusion. Thermal-visible fusion addresses this limitation by combining the textural richness of RGB imagery with the illumination-invariant heat signatures captured by long-wave infrared sensors. This paper presents a structured research study on pedestrian detection enhancement via thermal-visual fusion. The study has two goals: first, to synthesize the evolution of the field across benchmark datasets, fusion strategies, and deep architectures; second, to formulate a practical reliability-aware fusion pipeline suitable for student implementation and comparative evaluation. The paper organizes prior work from early aggregated channel features to attention-based and transformer-based multimodal detectors, and it analyzes why mid-level fusion, modality alignment, and illumination-aware weighting consistently outperform naive input-level concatenation. Quantitative evidence compiled from public KAIST, LLVIP, and FLIR benchmark reports shows that multispectral fusion reduces miss rate substantially compared with single-modality baselines and produces more stable day-night behavior. Building on those observations, a proposed architecture is described that integrates modality-specific encoders, geometric alignment, channel-spatial attention, confidence-guided feature fusion, and a lightweight one-stage detection head. The paper further outlines a reproducible evaluation plan, implementation parameters, failure taxonomy, and future research opportunities including infrared-centric reasoning, transformer fusion, uncertainty estimation, and edge deployment. The presented results are benchmark-oriented comparative findings synthesized from the literature and are intended to guide future experimental work rather than claim a new private leaderboard.
Registration ID: IJVRA_702416 Published ID: IJVRA2603771
"Pedestrian Detection Enhancement via Thermal-Visual Fusion", IJVRA - International Journal of Versatile Research and Analysis (www.IJVRA.org), ISSN:2984-8903, Vol.4, Issue 3, page no.948-958, March-2026, Available :https://ijpub.org/ijvra/papers/IJVRA2603771.pdf
Paper Reg. ID: IJVRA_702416
Published Paper Id: IJVRA2603771
Research Area: Other area not in list
Country: dondaicha, Maharashtra, India
ISSN: 2984-8903 | IMPACT FACTOR: 9.12 Calculated By Google Scholar | ESTD YEAR: 2023
An International UGC CARE JOURNAL PUBLICATION Low Cost (₹599), Scholarly Open Access, Peer-Reviewed, Refereed Journal Impact Factor 9.12 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage, Crossref DOI Member Journal Indexing in All Major Database & Metadata, Citation Generator
Publisher: IJVRA (IJ Publication) Janvi Wave
© 2026 - Authors hold the copyright of this article. This work is licensed under a Creative Commons Attribution 4.0 International License and The Open Definition.
This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0). 🛡️ Disclaimer:
The content, data, and findings in this article are based on the authors’ research and have been peer-reviewed for academic purposes only. Readers are advised to verify all information before practical or commercial use. The journal and its editorial board are not liable for any errors, losses, or consequences arising from its use.
: Click Here to Get
About Hard Copy and Transparent Peer Review Report
