An algorithm (COLMSTD) for detection of defects on rail and profile surfaces

• Published in: International journal of computer science and information security Vol. 14; no. 4; p. 45
• Main Authors: Orak, Ilhami Muharrem, Celik, Ahmet
• Format: Journal Article
• Language: English
• Published: Pittsburgh L J S Publishing 01.04.2016
• Subjects: Algorithms; Color; Computer information security; Computer science; Defects; Image detection; Rails; Rolling; Standard deviation; Surface defects
• ISSN: 1947-5500

Rail or profile products are used in many fields today. The rolling process is the most important production phase of the rail and the profile product. However, undesirable defects in the surface of the product during the rolling process can occur. Identifying these defects quickly by an intelligent system using image processing algorithms will provide a major contribution in terms of time and labor. For the detection of the regions, objects and shapes on the image, several algorithms were used. In this study, we introduce a Standard Deviation based algorithm (COLMSTD) by using the pixel color values. In order to evaluate the performance of the algorithm, the result of the COLMSTD algorithm is compared with the results of Hough Transform, MSER, DFT, Watershed, Blob Detection algorithms. In this study, it was seen that each algorithm has different capability in some extend to identify the surface defects in rail or profile. However, COLMSTD algorithm achieve more accurate and successful results than the other algorithms.