UAV facade inspection enhanced with aligned RGB-thermal fusion

Researchers have developed a new pipeline for aligning RGB and thermal imagery captured by unmanned aerial vehicles (UAVs) for facade anomaly detection. This system corrects for geometric discrepancies between sensors and enhances thermal image contrast to better identify surface and subsurface anomalies. A new dataset, M3T, containing 674 paired RGB and radiometric thermal samples, was introduced to evaluate the pipeline, which achieved a median registration error of 3.384 pixels and a mean average precision of 0.168 on specific anomaly detection tasks. AI

IMPACT Introduces a novel data fusion technique that could improve automated inspection systems for infrastructure.

RANK_REASON Academic paper detailing a new method and dataset for computer vision. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• DJI Mavic 3T
• RGB color model
• RGBT-Tiny
• unmanned aerial vehicle

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →