超声C扫描图像缺陷标记及边缘跟踪的研究

对数据量很大的C扫描图像进行缺陷成分标记时,传统的光栅扫描算法性能比较有限.为提高缺陷成分标记的效率,在第一次标记时利用辅助表记录等价标记号,标记完毕后去除重复记录、进行等价分析、补全记录和重新分配标记号,在第二扫描时修正C扫描图像缺陷标记号,试验证明了该方法的有效性.为获得边缘清晰而且轮廓完整的缺陷边缘,提出基于边界元运算的C扫描图像边缘跟踪算法,简化了跟踪过程.工程应用实例证明,在C扫描图像的缺陷边缘检测中该算法优于Gauss-Laplace算法和Prewitt算法.

Study on flaw labeling and edge tracking in ultrasonic c-scan image

Performance of traditional raster scanning algorithm is limited on labeling the massive data c-scan image. To improve efficiency of flaw component labeling, assistant table was used to save equivalent label at the first scanning, after some disposal processes to assistant table, such as wiping off repeated records, equipollence analysis, adding records and redistributing labels, C-scan image elements was relabeled at the second scanning, applications testify the validity of the method. To detect clear-cut and intact flaw edge, an algorithm based on edge element was proposed, simplifying the process of tracking edge, Engineering applications show that the algorithm is superior to Gauss-Laplace and Prewitt algorithms in c-scan image flaw edge tracking.