工业射线图像质量的自检测软件系统研究与实现

发布时间：2019-01-17 18:29

【摘要】：伴随着我国工业发展水平的提高,工业产品的需求和产出在日益增加,相应的工业产品质量控制难度也在加大,传统的检测过程大多依赖人工完成,在不增加人力投入的基础上,显然无法满足日益增加的检测需求,实现检测自动化是长远且有效的解决办法。无损检测技术是控制产品质量的一种有效手段,而射线检测技术是无损检测首选方法之一,随着技术的发展,射线数字成像技术将成为该领域应用的趋势。该技术可以形成被检物件的数字透照图像,同时数字图像支持计算机图像处理技术,计算机技术的应用为提高射线检测过程的自动化水平提供了可能。传统数字射线检测过程可以分为两步,首先需要对数字图像本身质量进行鉴定,然后从合格的数字图像中获取被检物件的缺陷信息。本文基于数字射线检测原理,解决的问题是如何使用计算机代替人工完成射线图像本身质量的鉴定,通过图像处理技术实现图像质量的自检测。首先,本文对传统射线图像质量的检测工艺进行了研究和介绍,从中总结得到图像质量评定方法和相关原理,并根据方法和原理制定了由自动获取阈值模块、滤波平滑模块和矩形识别模块组成的软件设计方案。其次,对各模块中涉及到的算法的原理进行研究,在此基础上针对本课题所涉及的问题对算法进行了适应性改进,并通过实验对改进后的算法表现进行评估,改进后的算法在运算效率方面有所提高。通过对三个功能模块的整合实现了射线图像质量自检测系统,使用该系统对大量实际透照图像进行检测,并与人工检测结果进行比较,完成了对系统可行性的测试。本文最后对所做工作进行了总结,并对未来工作进行了展望。通过大量实验表明,本文所提出的软件系统是有效可行的,基于图像处理技术的射线图像质量自检测功能在执行效率、判断精度和等方面相对于人工检测都具有优势,且具有很好的发展空间。
[Abstract]:With the improvement of the level of industrial development in China, the demand and output of industrial products are increasing day by day, the corresponding difficulty of quality control of industrial products is also increasing, the traditional detection process mostly depends on manual completion. On the basis of not increasing manpower input, it is obvious that it is unable to meet the increasing demand of detection, and it is a long-term and effective solution to realize test automation. Non-destructive testing (NDT) technology is an effective means to control product quality, and X-ray testing technology is one of the preferred methods of NDT. With the development of technology, digital radiography technology will become the trend of application in this field. This technique can form the digital radiographic image of the object under inspection, and the digital image supports the computer image processing technology. The application of the computer technology provides the possibility to improve the automation level of the radiographic detection process. The traditional digital ray detection process can be divided into two steps. Firstly, the quality of the digital image itself should be identified, and then the defect information of the subject object should be obtained from the qualified digital image. Based on the principle of digital ray detection, the problem solved in this paper is how to use computer instead of manual to complete the quality identification of the radiographic image itself, and to realize the self-detection of the image quality by image processing technology. First of all, the traditional radiographic image quality detection technology is studied and introduced in this paper, from which the image quality evaluation method and related principles are summarized, and the automatic acquisition threshold module is developed according to the method and principle. The software design scheme of filtering smooth module and rectangle recognition module. Secondly, the principle of the algorithm involved in each module is studied, on the basis of which the adaptive improvement of the algorithm is carried out, and the performance of the improved algorithm is evaluated through experiments. The improved algorithm improves the computational efficiency. Through the integration of the three functional modules, the self-detection system of radiographic image quality is realized. The system is used to detect a large number of actual radiographic images, and compared with the results of manual detection, the feasibility of the system is tested. At the end of this paper, the work done is summarized and the future work is prospected. A large number of experiments show that the software system proposed in this paper is effective and feasible, and the self-detection function of image quality based on image processing has advantages over manual detection in terms of execution efficiency, accuracy of judgement and so on. And has very good development space.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【参考文献】