基于ButterWorth滤波的X射线钢管焊缝缺陷检测方法

Automatic detection of weld defects in X-ray based on ButterWorth filtering

焊缝缺陷检测是保证焊接质量的重要环节,随着工业的高速发展和迫切需求,基于X射线焊缝缺陷自动检测技术得到了广泛的研究,但是由于成像方式,以及铸件材质等客观因素的影响,X射线图像存在噪声多、对比度低、背景亮度不均匀、焊缝边缘模糊等问题,使得利用计算机进行焊缝缺陷自动检测的准确率不太理想。针对这一问题,提出了一种基于ButterWorth滤波的钢管焊缝缺陷检测方法。首先利用傅里叶变换将图像的时域信息变换到频域,之后对图像的频域信息进行高通滤波,提取出图像中的高频缺陷信息,再对其进行傅里叶反变换,从而将缺陷信息凸出,最后利用阈值分割成功提取出缺陷,并对提取出的缺陷结果进行处理,有效地降低了漏检率和误检率。与其他传统检测算法相比,它对缺陷类型不敏感,具有较好的适应性和通用性。

Weld defect detection is a key link to ensure the quality of welding.With the rapid development of industry and urgent demand,the automatic detection technology of weld defects based on X-ray has been widely studied.However,because of the imaging methods,and the influence of casting material and other objective factors,X-ray image has many problems,such as noise background,low contrast,brightness uneven,and weld edge blur,which make the automatic detection accuracy of weld defects is not very ideal.For this problem,a new method,the weld defect detection algorithm based on ButterWorth filter,is proposed in this paper.Firstly,it uses the Fourier transform to transform the time domain information of the image into the frequency domain,then carries on the high-pass filtering to the frequency domain information of the image,and extracts the high frequency information of defects in the image,then transforms it in inverse Fourier transform,and highlights the defect information.Finally,using the method of threshold segmentation to extract the defects successfully,and further processing on the extracted results to reduce the false detection rate.Compared with other traditional detection algorithms,it is not sensitive to defect types,so it has better adaptability and versatility.