Detection of internal corrosion by long-pulse thermography and digital image processing

Abstract

This paper addresses the problem of internal pitting corrosion in metals. The approach to solving this problem is carried out by means of long-pulse thermography (step heating) and digital image processing, using the median filter in the pre-processing and the Fourier transform in the processing. The long pulse thermography technique is an excellent way of detecting corrosion in metals. According to the state of the art, most of the works reviewed detect only the presence of external corrosion. In the experiments plates of different thickness were evaluated, one of 2 mm and another of 8 mm; this is another point in favor since most of the reviewed articles only focus on a single sample. In the development of this project, the following lines of research have been identified: implementation of new digital image processing methods, improvement of the thermography technique using other energy sources to heat the plates in a shorter time, exploring other functions of image processing in programming languages, carrying out tests with plates of different thickness and size, measuring the percentage of corrosion.