非破壊検査・診断Ⅳ

[1292] A CRACK DETECTION METHOD FOR CONCRETE INFRASTRUCTURES BASED ON IMAGE PROCESSING TECHNIQUE AND GENETIC ALGORITHM

CUONG NGUYEN¹, 河村圭¹, Amir TARIGHAT², 中村秀明¹ (1.山口大学 2.Shahid Rajaee Teacher Training University)

キーワード：genetic algorithm, image processing, crack detection, concrete infrastructure

This paper presents a crack detection method based on genetic algorithm to optimize the parameters of image processing technique (IPT). Morphological mathematic is applied to the pre-processing step in order to smooth the original image. Next, the global binarization and dilation-thinning operation is used for segmentation and noise removal. Moreover, the experimental results of the proposed method compared with other methods for concrete infrastructure surface images under various complex photometric conditions validate the reasonable accuracy in practical applications.