@Article{cmc.2023.027102, AUTHOR = {Fengyu Xu, Masoud Kalantari, Bangjian Li, Xingsong Wang}, TITLE = {Nondestructive Testing of Bridge Stay Cable Surface Defects Based on Computer Vision}, JOURNAL = {Computers, Materials \& Continua}, VOLUME = {75}, YEAR = {2023}, NUMBER = {1}, PAGES = {2209--2226}, URL = {http://www.techscience.com/cmc/v75n1/51417}, ISSN = {1546-2226}, ABSTRACT = {The automatically defect detection method using vision inspection is a promising direction. In this paper, an efficient defect detection method for detecting surface damage to cables on a cable-stayed bridge automatically is developed. A mechanism design method for the protective layer of cables of a bridge based on vision inspection and diameter measurement is proposed by combining computer vision and diameter measurement techniques. A detection system for the surface damages of cables is de-signed. Images of cable surfaces are then enhanced and subjected to threshold segmentation by utilizing the improved local grey contrast enhancement method and the improved maximum correlation method. Afterwards, the data obtained through diameter measurement are mined by employing the moving average method. Image enhancement, threshold segmentation, and diameter measurement methods are separately validated experimentally. The experimental test results show that the system delivers recall ratios for type-I and II surface defects of cables reaching 80.4% and 85.2% respectively, which accurately detects bulges on cable surfaces.}, DOI = {10.32604/cmc.2023.027102} }