Using Image Processing to Find Weld Defects in Radiography Films

Abstract

The use of image processing is becoming more widespread in a variety of industries, including robotics, agriculture, medicine, and meteorology. Numerous investigations have been carried out in these domains; however, there has been minimal research on its utilisation in weld inspection. The groove and complete joint piercing high-strength failures in welding (such as pressure vessels, heat boilers, etc.) were examined using the radiographic testing method. Image processing can be used to optimise images and minimise or remove faults in radiography films when there are similar flaws with variable acceptance criteria. Edge detection, better picture quality, and precise colour diagnosis are all feasible in image processing, and they all aid in precisely identifying defects and reducing errors in defect type diagnosis. This project will look into the process of employing image processing to detect internal weld faults in radiography films. The goal is to fully automate the process, doing away with the requirement for human interpretation of the film. First, it will be important to discuss both the various types of weld imperfections and the general and fundamental notions of image processing in order to efficiently identify defects using image processing approaches. Subsequently, using the findings of research and development, procedures and methods for implementing these algorithms will be explained, along with an explanation of how to use MATLAB software.