摘要
刻槽能有效地增加跑道道面摩擦力和排水能力,提高机场跑道的服务性能和安全性.由于三维结构和表面纹理的差异,传统基于二维可视化技术的刻槽病害检测方法难以满足刻槽几何参数高精度检测要求.提出了一种基于三维线扫描传感器数据的道面刻槽检测方法.利用三维线扫描传感器获取跑道各断面的数据,建立沟槽断面数据库.根据贝叶斯分类原则和局部相关耦合约束条件,在断面尺度上对刻槽进行分割和分类.根据刻槽空间纹理评价的回归条件,对完整的刻槽进行识别和测量.香港国际机场的实验结果表明,该方法具有良好的识别精度和泛化性能.此外,实验还表明,正常刻槽适用于形状特征分类,而病害刻槽,尤其是严重疾病刻槽适用于纹理特征分类.
The grooves effectively increase the friction and drainage capacity of runway surface,improve the performance and safety of runway.Because of the differences of structure and content,it is still a difficult problem to achieve the precision identification of grooves geometric based on the 2D measurement technologies.A detection method of grooves on the runway surface based of the data of 3D line scanning sensor is proposed.The data of each profile of the runway is obtained by 3D line scanning sensors,and the database of grooving profile is made.The grooves are segmented and classified at the profile scale according to the Bayes classification and the local correlation coupling constraint condition.The complete grooves are identified and measured according to the regression condition of the groove space texture assessment.The experimental results of Hong Kong International Airport show that the proposed method has good accuracy and generalization performance in groove recognition and evaluation.Further experiments also suggest that normal grooves are suitable for shape feature classification,while disease grooves,especially severe ones,are suitable for texture feature classification.
作者
张德津
王毅
章洋
汤青洲
田霖
ZHANG Dejin;WANG Yi;ZHANG Yang;TANG Qingzhou;TIAN Lin(School of Electrical and Electronic Engineering, Hubei University of Technology, Wuhan 430072, China;Wuhan Wuda Zoyon Science and Technology Co., Ltd., Wuhan 430223, China)
出处
《中南民族大学学报（自然科学版）》
CAS
2020年第2期195-204,共10页
Journal of South-Central University for Nationalities：Natural Science Edition
基金
国家自然科学基金资助项目(51379081)。
关键词
机场道面
刻槽检测
三维断面
机器学习
airport runway
grooves detection
3D laser profiling
machine learning
