Color Composite of Digital Radiography Image and Region Growing Defect Detection

Abstract:

DR image in industrial castings is usually a gray scale image and it has difficulty in simultaneously showing the defects within different thickness parts. To address this issue, the floating point type data obtained by the scanning of DR system is firstly converted into three gray scale images by the more gray stretch technique, where different gray scale images contain different thickness defects of the castings. Secondly, those three gray images are synthesized into one color image. The synthesized image contains lots of details of measured castings and its aim is to reduce the loss of image detail caused by the converting procedure. Then region growing model is exploited to segment the defects of images. Finally, experimental results show that the color synthesis method can better segment the casting defects of the different wall thickness.

• Periodical: Applied Mechanics and Materials (Volumes 380-384)
• Pages: 938-942
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.380-384.938
• Online since: August 2013
• Authors: Kai Tian, Jia Rong Shi
• Keywords: Color Composite, Defects Detection, Digital Radiograph (DR) Image, Image Segmentation, Region Growing

References

[1] Robert A M, Sunil K K: Trenchless Technol Res (1999).

[2] Gongtian Shen, Lifei Li: Non-Destructive Testing In Chinese (2011).

[3] Blakeley B: INSIGHT (2004).

[4] Bueno C, Barker M D: X-Ray Detector Physics Applications (2009).

[5] Adams R, Bischof L: IEEE Transactions on Pattern Analysis and Machine Intelligence (1994).

[6] Frank Y: Image and Vision Computing (2005).

[7] Rafael C. Gonzalez, Richard E. Woods: Digital Image Processing (Electronics Industry Publications, Beijing 2009).

[8] Kai Tian, Li Zeng and Linghui Liu: Computer Engineering In Chinese (2011).