Development of software to interpretate radiographs of F5 wing spars focussing on crack recognition

• Main Author: Seah, Daphne Xinrui.
• Other Authors: School of Mechanical and Aerospace Engineering
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Aeronautical engineering
• Online Access: http://hdl.handle.net/10356/54026
• Institution: Nanyang Technological University

The Republic of Singapore Air Force (RSAF) is developing an automated digital radiography system which comprises an automated crack detection software, to identify cracks on the F5 wing spars. This development will serve to ease the workload of inspectors in their visual and manual crack detection routine especially when it comes to complicated radiographs, and through this also reduce human errors and inaccuracies that could come with it. For this project, we capitalized on the idea of improving on the existing software developed by previous Final Year Project students and specialists. Analysing the existing literature will give an idea on the various problems faced and unsolved, and how the codes can be improved to meet the objectives of this project. It is also of priority to enhance the robustness of the program. The new algorithm saw major alterations made to certain segments of the original one. A new approach of using pixel grey levels to identify features on the radiographs was adopted. Results obtained were generally encouraging and positive. For future developments, more real crack images should be put to test to enhance robustness of the program, and focus should go into analyzing simpler parts of a complicated radiograph by further segmenting them, so that inspection can be done with greater precision. The success of the crack detection software largely depends on the accuracy of control on radiography environment, and more importantly a larger pool of test images has to be available to the program developers, as all these will contribute to creating a more robust program.