Radiographic DS Level Imaging And Weld Defect Database Construction System

The welding quality of oil and gas pipeline construction is related to engineering safety.Digital conversion and intelligent evaluation of radiographic film is a research hotspot in the current NDT industry,but it mainly faces the following three problems :（1）the current way of manually controlled camera shooting is inefficient and the image quality is unstable;（2）The digital radiographic imaging quality can’t meet the relevant national standards in the industry,namely,GB/T26141 digital radiography negatives the nondestructive testing system of quality evaluation "of the DS level,lead to not be able to use digital image for the follow-up of the negative and storage management,still need to rely on physical film cause inefficient effort in evaluation and management;（3）In actual industrial production,the number of defect image samples serving for intelligent defect detection is seriously insufficient,and there is no open source standard weld defect database.In this paper,based on the requirement of national standard,combining of radiographic examination in the process of industrial production processes,use QT framework to design a set of radiographic DS weld defect imaging and build a system,realize the radiographic examination to meet the requirements of the DS level indicators,digital imaging and build up meet the demand of defect detection data library of weld defect.The main research contents of this paper are as follows:（1）Realization of radiographic automatic exposure imaging moduleThis module aims to target different blackness of radiographic complete digital transformation,first of all,according to the analysis of the research object,designed a kind of automatic exposure algorithm based on the statistical characteristics of the image,by acquiring the photographed image statistical characteristics of the pixel,to quantify the degree of exposure,exposure again according to the different situation to adjust the parameters of different amplitude,The function of automatically adjusting exposure parameters according to each film with different blackness is realized.A multi-level caching mechanism is designed for different lengths of negative images.Finally,the standard film required in the NATIONAL standard is used to verify the indicators of the DS level.（2）Realization of DCGAN network amplification moduleFor the digital transformation of negative images,images with defects were selected,and local images containing defects were made into original data sets.An image data amplification method based on the DCGAN network was designed.After setting up the network model,a large number of newly generated defect samples were obtained through parameter adjustment and iterative training.Then the generated local defect image is fused to different positions of the full negative image to obtain the final amplified image sample.（3）Realization of construction module of weld defect databaseFor lack of data sets and defect detection,non-standard problems,using the amplification with defects generated negative images,designed the weld defect with the function of radiographic image enhancement library building software,image enhancement algorithm is used to analyze the defect information enhancement,can be more accurate judgment and annotation,eventually to save defect information into XML format file,Thus it provides rich data support for subsequent defect detection.After software testing and practical trial,the DS radiographic imaging and weld defect database construction system designed in this paper has perfect functions and runs stably.It can complete the adaptive exposure imaging of the film with a blackness range of 0.5-4.5,and its imaging quality meets the requirements of DS level in the national standard.When using this system to shoot,the process is efficient and the results are accurate.The images taken can not only be used as a part of the weld defect database,but also as the data source of various subsequent intelligent testing and operation of the project,which has good application value and market prospect in the field of nondestructive testing.