一种基于超声检测图像的缺陷自动识别算法

A method for automatic defect recognition based on ultrasonic image

传统探伤都是通过大量的人力对探伤图逐一判断,探伤效率低下,正确率不高。本文针对探伤A超图像序列提出了一种自动识别探伤的算法。通过分析探伤图像,首先对原图进行k-means聚类的分割,得到带有虚景的探伤声波图像。为了抑制虚警,得到完整的声波图像,本文使用了投影算法,并取得了很好的效果。最后在得到的声波图像上进行底波和缺陷波的检测,实现缺陷的自动识别。实验表明,本文提出的方法具有很高的准确率。

Traditional detection is through a lot of human judgment for defect recognition. The efficiency is low and the accuracy is not high. In this paper, we put forward an automatic recognition algorithm on the defect image sequence. After analyzing the defect image, first of all, we apply k-means cluster segmentation oil the original image to get the acoustic images with false alarm. In order to get full acoustic images by preventing the false alarm, we use the projection algorithm and achieved good results. Finally we detect the bottom wave and defect wave on the acoustic image, and realize the recognition of defects automatically. The experimental results show that the proposed method has high accuracy.