Investigation of Corrosion Intensity on Metal Structures Using Image Processing Algorithms

Abstract

An investigation of corrosion intensity on metal structures using corrosion detection algorithms is one of the advanced methods for corrosion detection since the traditional method is time-consuming for vast areas and impossible to investigate in inaccessible areas. Estimation of the intensity of the corrosion by humans may not have much accuracy, and false analysis of the metal surfaces may lead to the destruction or loss of the material. To cater to the above needs intelligent corrosion detection algorithms are developed, which automatically detect and inspect the corrosion on metal surfaces by using a series of photographic images. percentage of black pixels is one of the parameters for analyzing the corrosion intensity. The first algorithm initially converts the RGB image to a grayscale image and it is binarized using Otsu's method, and from the binary image, PBP is calculated. The higher the value of PBP higher is the corrosion intensity. Another approach made to investigate the corrosion is by verifying the roughness and smoothness of the metal surface since a rough surface indicates the presence of corrosion. This is achieved by calculating the energy of the image, which is one of the properties of GLCM.