论文部分内容阅读
基于X射线成像的焊缝缺陷自动检测技术对提高工业射线检测的自动化水平具有重要意义。焊缝缺陷在线连续检测的实时性要求较高,随着工件厚度的增加,其焊缝X射线实时图像的信噪比变得很低,使得已有的处理算法难以在满足实时性的同时,有效处理缺陷误检与漏检之间的矛盾。针对这些问题,在分析了传统方法在厚壁工件X射线图像焊缝缺陷自动检测中存在的问题基础上,对传统方法进行了改进,提出了双阈值背景消除法和平行焊接方向波形分析法,然后利用所提出算法之间的冗余性和互补性,融合多种分割结果以解决缺陷误检与漏检之间的矛盾。试验结果表明:所提出的缺陷自动检测方法能够在满足实时性要求的同时,实现缺陷检出,有效避免误检。
The automatic detection of weld defects based on X-ray imaging is of great significance to improve the automation level of industrial radiography. The real-time requirement of on-line continuous inspection of weld defects is high. With the increase of workpiece thickness, the signal-to-noise ratio of real-time image of weld X-ray becomes very low, making the existing processing algorithms difficult to meet the real- Effectively deal with the contradiction between the defect and the missed inspection. Aiming at these problems, the traditional method was improved based on the analysis of the existing problems in the automatic detection of weld defects of X-ray images of thick-walled workpieces. The double threshold background elimination method and parallel welding direction waveform analysis method were proposed. Then, by using the redundancy and complementarity between the proposed algorithms, a variety of segmentation results are fused to solve the contradiction between the false positives and negative ones. The experimental results show that the proposed automatic defect detection method can meet the real-time requirements and detect defects at the same time, effectively avoiding false detection.