复杂道路场景下无人机航拍视频去雾方法研究OA

To address the challenge of dehazing unmanned aerial vehicle(UAV)videos cap-tured in complex road environments,this study constructed a video dataset using the DJI Mini 4 Pro covering various scenes along Yunxiang Avenue and Huixian Avenue in Nan-chang.Three representative video dehazing methods-based on the Clearplus plugin,Dark Channel Prior and Retinex theory-were comparatively analyzed.Both objective metrics,in-cluding Peak Signal-to-Noise Ratio(PSNR)and Structural Similarity Index(SSIM),and a professional subjective visual evaluation framework were employed for comprehensive as-sessment.Experimental results indicate that the methods exhibit distinct performance char-acteristics under complex road conditions:The Clearplus-based approach achieves the best overall performance;the Retinex-based method performs exceptionally well in dense fog conditions;and the Dark Channel Prior method excels in preserving fine image details.This study provides a quantitative reference for selecting optimal dehazing solutions in practical applications such as intelligent traffic monitoring and UAV-based road surveillance.