Estimation of Metal Surface Defect Using CD Segmentation

Analysing metal surfaces of massive metal instruments like boilers are tedious process, which also require very specialized equipments. To employ the surface analysis schemes at a lower cost, image processing facilitates many processing methodologies. Using the camera captured live images of the metal surface from the massive metal equipments, image processing techniques can be applied and similar results can be obtained as like the specialized analysing equipments. To improve the image processing techniques, series of methodologies are proposed and implemented to analyse the metal surface image for any defect found. The proposed system employs noise removal, false negative removal followed by segmentation using the proposed clustering method. The results obtained by the proposed system are encouraging and are comparable with other existing image processing methodologies.