科技创新与应用2025，Vol.15Issue(31)：1-7,7.DOI:10.19981/j.CN23-1581/G3.2025.31.001
基于自适应边缘检测与Scharr算子的踏面剥离检测
摘要
Abstract
Aiming at the problem that the measurement error of the wheelset measuring machine increases due to the inconsistency of the wheelset axle diameter and wheel diameter,the measuring machine is improved by adding an image detection device to assist in judging whether the wheelset is raised in place,and the wheelset tread defect detection based on image recognition is carried out.Adaptive median filtering and adaptive gamma transform algorithm were used to improve image preprocessing.The traditional edge detection algorithm was optimized by introducing the Otsu algorithm to adaptively determine thresholds and adopting the Scharr operator for gradient calculation,which strengthened gradient responses in diagonal directions.The test results show that the added image detection device improves the accuracy of wheel set lifting position,and the established image recognition algorithm has excellent performance in denoising,image enhancement,edge detection and damage recognition,which effectively overcomes the defects of traditional algorithms and significantly improves the accuracy and reliability of wheel tread peel recognition.It provides more efficient and reliable technical support for train wheelset tread peeling detection.关键词
轮对测量机/图像识别/车轮踏面/剥离识别/边缘检测Key words
wheelset measuring machine/image recognition/wheel tread/peeling recognition/edge detection分类
计算机与自动化引用本文复制引用
李兴城,方智超,王飞..基于自适应边缘检测与Scharr算子的踏面剥离检测[J].科技创新与应用,2025,15(31):1-7,7.基金项目
国能铁路装备有限公司科研项目(2023-026) （2023-026）
