Recognition of Concrete Surface Cracks Using FCM-based RBF Network

This paper, first, uses the closing morphologic operation to adjust the effect of light extending over the whole image of concrete surface. After applying the high-pass filtering operation to sharpen boundaries of cracks, this paper classifies intensity values of the image to 8 groups, removes intensity values affiliated with a group having the highest frequency among 8 groups for the removal of background, and binarizes the preprocessed image. The assistant lines used to measure cracks of concrete surface are removed from the binarized image using position information extracted by the histogram operation, and cracks broken by the removal of background are extended to reconstruct an original crack by 5×5 masking operation. This paper removes unnecessary information by applying three types of noise removal operation successively and extracts areas of cracks from the binarized image. At last, the opening morphologic operation is applied to compensate extracted cracks and characteristics of cracks are measured on the compensated ones. We proposed the method for automatically recognizing the directions of the cracks with the FCM-based RBF network. Experiments using real images of concrete surface showed that the proposed method extracts cracks well and precisely measures characteristics of cracks. Also, The proposed FCM-based RBF network was effective in the recognition of direction of the extracted cracks.