论文部分内容阅读
针对油气管道X射线底片传统评片方式效率低、检测结果主观性大等原因，在X射线底片数字化的基础上利用图像特征提取的方法实现管道缺陷自动识别。由于X射线底片图像具有噪声大、对比度低等特点，又因其缺陷面积、圆度及周长等几何特征值与边界轮廓有关，本文提出一种改进Canny边缘检测算法对缺陷的边缘轮廓进行提取。实验结果表明，本文方法能有效实现对缺陷特征的提取，且实时性强。“,”The traditional artificial rating on X -ray film of oil pipeline is of low efficiency ,and automatical defects recognition can be real-ized on digitalized X-ray film by feature extraction .But the X-ray film image is characterized by great noise and low contrast ,in which the value of area ,roundness and perimeter are related to the boundary contour ,an improved canny algorithm on defects is proposed for edge detection .The experimental results demonstrate that the proposed approach can better realize the feature extraction of defects with real -time effect .