Abstract: In many manufacturing industries, conventional defect detection is done by experienced human inspectors who sketch the defect patterns or defect types manually and then resolve them. However, such defect detection methods are more costly, less accurate, time consuming and complicated also. To overcome these problems, a new method has been developed to detect and identify the defects in industrial pipes, effectively as well as automatically. This method is based on image processing and embedded system. This method works on video input. This proposed method basically works in three steps. Firstly, it converts the RGB images of the pipe into a grayscale image. Secondly, it extracts the pipe as well as fault. Then lastly it detects and identifies the defect. It also gives the idea about major and minor defect. So, according to this well known decision of pipe modification or removal of pipe from process can be taken.

Keywords: Defect detection, defect identification, embedded system, image processing, pipe industry.