Publications / 2000 Proceedings of the 17th ISARC, Taipei, Taiwan

Application of Neuro-Fuzzy Thresholding to Bridge Painting Defects Recognition

Po-Han Chen, Luh-Maan Chang
Pages 1-6 (2000 Proceedings of the 17th ISARC, Taipei, Taiwan, ISBN 9789570266986, ISSN 2413-5844)
Abstract:

The neuro-fuzzy thresholding method segments a grayscale image into three areas in accordance with the illumination values of the pixels in the image. The three average illumination values of the three areas are then sent to a pre-trained neural network to obtain the three corresponding threshold values. Fuzzy theory is used to adjust the gray level values of the pixels along the boundaries between areas. Finally, image thresholding is applied to get a binary image containing only the object pixels and the background pixels.

Keywords: neural networks, fuzzy adjustment, image thresholding, k-means algorithm, neurofuzzy thresholding.