基于傅里叶变换的焊缝表面裂纹图像识别技术研究
Research on Image Recognition Technology of Weld Surface Cracks Based on Fourier Transform
梁小虎1
作者信息
- 1. 天津市特种设备监督检验技术研究院 天津 300192
- 折叠
摘要
为提高焊缝表面裂纹的检验效率,有效降低检验成本和高空坠落风险,本文将图像识别技术应用于焊缝表面裂纹识别.针对自然条件下焊缝与图像背景对比度低的特点,提出了基于二维傅里叶变换的焊缝识别技术,实现了对焊缝方向和焊缝宽度的识别,有效降低了图像背景对裂纹识别结果的影响,提高了对焊缝表面裂纹的检出率.通过图像识别实验验证了本文提出的方法可以实现对焊缝表面裂纹的快速、有效筛查.
Abstract
In order to improve the inspection efficiency of weld surface cracks and effectively reduce the inspection cost and the risk of high-altitude fall,this paper applies image recognition technology to weld surface cracks recognition.Aiming at the characteristics of low contrast between weld and image background under natural conditions,a weld recognition technology based on two-dimensional Fourier transform is proposed,which realizes the recognition of weld direction and weld width,effectively reduces the influence of image background on crack recognition results,and improves the detection rate of weld surface cracks.The method proposed in this paper has been validated through image recognition experiments to achieve rapid and effective screening of surface cracks in welds.
关键词
图像处理/焊缝/表面裂纹/图像识别Key words
Image processing/Weld/Surface cracks/Image recognition引用本文复制引用
出版年
2024
NETL
NSTL
万方数据