Implementation of Computer Vision Technique for Crack Monitoring in Concrete Structure

摘要

Assessment of structural health is essential for safe and efficient functioning of built environment. Physical inspection of structures for its health monitoring is time-consuming, costly and risky. Advances in image acquisition, processing techniques, and computational resources have made computer vision a cost effective and an accurate technique for structural health assessment. Recent evolution of Convolutional Neural Network has reduced human effort and made it easy to develop algorithms for identification of structural defects. One of the primary defects in concrete is crack. Concrete cracking occurs due to many reasons like shrinkage, heaving, premature drying, excessive loading etc. and it leads to reduction in strength of structures. This paper presents a computer vision system developed for crack monitoring of concrete cubes subjected to compressive loading. Camera is used to capture real-time images when concrete cubes are subjected to loading. Images are processed using the convolutional neural network to identify crack and subsequently features of cracks like number, location, length, and area are extracted. The outcome of present system demonstrated better and accurate real-time monitoring of cracking when concrete is subjected to loading. The proposed computer vision-based approach is a step forward in Structural Health Monitoring of real-life concrete structures like buildings, bridges, and pavements.