首页|基于图像处理的轨道表面病害检测研究

基于图像处理的轨道表面病害检测研究

扫码查看
以钢轨为研究对象,结合图像处理技术,对钢轨表面质量进行检测。首先,对钢轨图像进行预处理,通过线性灰度变换对图像进行增强,并采用自适应滤波进行图像去噪;其次,采用Ostu阈值分割算法对预处理后的钢轨图像进行缺陷分割;最后,以离心率、矩形度和致密度作为分类依据,通过决策树分类法对钢轨表面缺陷进行分类。实验结果表明:基于图像处理的钢轨表面质量检测方法能够有效对钢轨缺陷进行检测,分割准确率为 96。7%,分类准确率为90%,为钢轨表面质量检测提供了一种有效的检测方法。
Research on Orbital Surface Disease Detection Based on Image Processing
Taking steel rails as the research object,combined with image processing technology,the surface quality of steel rails is detected.Firstly,it preprocesses the steel rail image,enhance the image through linear grayscale transformation,and use adaptive filtering for image denoising.Secondly,the Ostu threshold segmentation algorithm is used to segment defects in the preprocessed steel rail images.Finally,using eccentricity,rectangularity,and density as classification criteria,the surface defects of steel rails are classified using decision tree classification method.The experimental results show that the image processing-based steel rail surface quality detection method can effectively detect steel rail defects,with a segmentation accuracy of 96.7%and a classification accuracy of 90%,providing an effective detection method for steel rail surface quality detection.

steel rail defectimage processingedge detectionthreshold segmentationdefect classification

胡璐萍、王琪璇、吴哲、王小龙

展开 >

西安交通工程学院 机械与电气工程学院,陕西 西安 710300

钢轨缺陷 图像处理 边缘检测 阈值分割 缺陷分类

陕西省教育厅科研项目

23JK0531

2024

现代信息科技
广东省电子学会

现代信息科技

ISSN:2096-4706
年,卷(期):2024.8(7)
  • 22