【作者】 周晋锋；

【导师】 王明泉；

【作者基本信息】 中北大学 ， 信号与信息处理， 2009， 硕士

【摘要】 随着电子技术和射线无损检测的发展,X射线实时成像设备的到了很大的发展。而X射线平板探测器实时成像技术现在成了研究开发的热点。但是由于平板探测器价格昂贵,国内相关研究和应用比较少。课题以PaxScan2520V平板探测器作为研究对象,针对药柱疵病X射线检测,配合厂家搭建了硬件系统,并开发了X射线实时成像系统软件。在分析了射线检测的物理基础与X射线平板成像的原理的基础上,设计了药柱疵病检测的射线实时成像系统。并对实时成像系统各项指标做了评价,评价结果满足检测需求。论文对平板成像系统中影响成像质量的主要环节进行了分析研究,对该过程的偏差、像元响应不一致及坏点等图像降质因素进行了有效抑制。对系统成像中图像随机噪声进行了校正并改进了工程化帧积分算法。缺陷的特征提取为智能识别和评判提供数据,为实现X射线检测的自动化提供理论基础。论文针对药柱X射线图像,总结以往缺陷特征提取的方法的基础上,引进模糊理论,利用了广义多层次模糊算子增强算法,实现了对药柱X射线图像缺陷边缘特征提取,实验结果证明,对药柱一般缺陷提取效果良好,并为以后药柱X射线缺陷自动识别打下良好的基础。综上所述,X射线实时成像系统达到了课题和厂家的要求。与其他成像系统相比有其优势和独特之处。但是还有补足,需要进一步完善。更多还原

【Abstract】 With the development of the electron-ray technology and non-destructive testing,X-rayreal-time imaging equipment has been greatly improved. Research and development ofreal-time imaging technology through X-ray flat panel detector has now become a hot spot.However, due to its high cost, the flat panel detector has not been widelystudied and appliedin China. This essay focuses on the study of PaxScan 2520V panel detctor which is used todetect the flaws in the grain through X-ray. On the request of the customers, the releventhardware system and real-time X-ray imaging system are also developed.It studies the physical basis of x-rayd etection and the X-ray flat panel detector imagingprinciple and designs the real-time x-ray imaging system used in grain flaw detection, andevaluates the system’s various character tables,all of which meet the detecting demand.The paper details the system’s image calibrating and noise reducing function, putsdeviation of the process, inconsistent pixel response and the Dead and other major elementswhich affect the imaging quality under effective control, corrects the random noise inimagingandimprovestheengineeringframeintegralalgorithm.Defect feature extraction provides datafor intelligent identificationand evaluation,andis the theoretical foundation of automatic X-ray inspection. This paper takes a grain’s X-rayimage as a study case, sums up the existing defect feature extraction technology, introducesthefuzzytheory, uses the generalizedmulti-level fuzzyoperatorenhancement algorithms andachieves edgeX-rayimagefeatureextractionwhichhasbeenprovedeffectiveinexperiments,layingasolidfoundationforbetterautomaticX-raydefectidentificationofthegrain.In summary, the X-ray real-time imaging system in this paper has met both theoreticaland practical demand .It enjoys great advantage compared with other imaging systems, but still with room of further improvement.更多还原