Abstract
Weathering steel has been used for structural components of steel bridges. This is because the steel obviates the need for painting. However, in some rare cases, the durability does not live up to expectations due to sever surface corrosion. Therefore, it is very important to evaluate rust condition for the proper maintenance. This paper presents a rust pattern classification method. The classifier was developed by neural networks. Then, three feature extraction methods, Fourier Transform, Wavelet Transform and Gray Level Co-occurrence Matrix (GLCM), were applied as inputs to the classifier.
