基于SURF算法的射线底片拼接技术研究

Research on image stitching of radiographic technology film based on SURF stitching algorithm

在工业射线技术中,由于其本身技术特点的原因以及射线底片长度有限,一次检测焊缝的长度使得检测往往需要多张底片彼此搭接在一起才能完整显示原有的信息。这样不仅影响了检测后的直观性,还存在着伪造焊缝射线底片的可能。目前,现有的防伪手段之一是由监督检测人员对射线底片进行人工识别,但是由于同一铸件的底片数量较多以及特征不明显等原因,存在着人工识别率和准确率较低的隐患。因此,针对上述问题,采取一种基于SURF算法的射线底片拼接技术,对射线底片的重叠区域进行拼接,将其拼成一幅完整的射线图像,分别对含有缺陷和不含缺陷的射线底片图像进行实验,结果表明,基于SURF算法可以很好的匹配底片重叠区域的特征点,能较快速的实现射线底片的无缝拼接,实现100%的拼接效果,此外,该研究实现的结果还能对底片快速辨识和避免射线底片的伪造现象的发生。

In industrial radiation technology,due to its own technical characteristics as well as the limitation of the length of the radiation negatives,the detection of the radiation negatives of the same weld often requires multiple negatives to be lapped together in order to display the original information completely,which not only reduces the intuitiveness of the detection results,but also has the risk of the radiation negatives being forged.At present,one of the existing anti-counterfeiting means is manual identification of the ray negatives by the supervisory inspectors,but due to the large number of negatives of the same casting,the features are not obvious and other reasons,leading to the hidden danger of low manual identification rate and accuracy.Therefore,in order to solve the above problems,we adopt a technical research based on SURF X-ray negative splicing technology to splice the overlapping area of the ray negative into a complete ray image,and carry out experiments on the ray negative with and without defects respectively,and the results show that the SURF algorithm based on SURF algorithm can match the feature points of the overlapping area of the negative well,and can quickly realize the ray.In addition,the results achieved by this method allow for quick identification of the negative and avoid counterfeiting of X-ray negatives.