光学精密工程2025，Vol.33Issue(15)：2414-2423,10.DOI:10.37188/OPE.20253315.2414
道岔参数在线测量中激光条纹中心自适应提取
Adaptive extraction of laser stripe center for on-line measurement of turnout parameters
摘要
Abstract
The accuracy of laser stripe center extraction critically determines the precision of structured-light vision measurements.In an online structured-light turnout-parameter measurement system,varia-tions in stripe position due to turnout movement and environmental disturbances produce uneven stripe brightness and width,degrading measurement accuracy.To address these issues,an adaptive extraction method for the stripe center is proposed based on an improved Steger algorithm,offering enhanced envi-ronmental robustness and high precision.The method first isolates the effective stripe region using Otsu thresholding combined with a Two-Pass algorithm to suppress stray-light interference and accelerate com-putation.The stripe skeleton is then extracted and initial center points are refined.A neighborhood geo-metric normal-direction computation is introduced to determine the stripe cross-sectional width.By corre-lating the measured stripe width with the Gaussian kernel parameter σ,the convolution-kernel size is adap-tively adjusted,yielding an accurate sub-pixel centroid of the stripe.Evaluation on a standard calibration block demonstrates an average measurement error within 0.09 mm.The proposed adaptive approach ex-hibits strong anti-interference capability and high accuracy,making it suitable for online measurement of turnout parameters.关键词
视觉测量/线结构光/中心线提取/Steger/自适应Key words
vision measurement/line structured light/centerline extraction/Steger/adaptive分类
信息技术与安全科学引用本文复制引用
荆颖春,何启欣,崔建英,张政,冯其波..道岔参数在线测量中激光条纹中心自适应提取[J].光学精密工程,2025,33(15):2414-2423,10.基金项目
国家自然科学基金重点项目(No.51935002) （No.51935002）
中央高校基本科研业务费专项资金资助项目(No.2023JBMC024) （No.2023JBMC024）
