Deep learning based semantic segmentation for cracking damage detection
Semantic segmentation based on deep learning has been applied to structural health monitoring(SHM)in recent years.Meanwhile,the traditional crack detection algorithms are being plagued by problems such as non-uniform illumination conditions,interference of noncracking information,discontinuous cracks and so on.This paper presents a new model based on the improvement of the original Deeplab V2 to accommodate the particularity of crack damage detection.The results show that the method can yield outputs with very high precision when compared with the previous methods,which would be favourable for cracking width measurement.The presented approach is promising to meet the engineering requirements for crack damage detection and provides some feasible suggestions for the application of deep learning in structural health monitoring.
crack damage detection deep learning semantic segmentation neural network
Yu Wang Jinyang Fu Junsheng Yang Kai Zhang Zhiheng Zhu
School of Civil Engineering,Central South University,Changsha,China School of Civil Engineering,Central South University,Changsha,China;National Engineering Laboratory
国际会议
The 7th World Conference on Structural Control and Monitoring(7WCSCM)(第七届结构控制与监测世界大会)
青岛
英文
1665-1672
2018-07-22（万方平台首次上网日期，不代表论文的发表时间）