Image Based Crack Recognition on Concrete Structures Using Convolutionalal Neural Networks

The objective of the study is to automate the Structural Health Monitoring and site inspection work by developing and modeling an Artificial Neural Network with Convolutional Neural Network technology using images of concrete cracks as data. The feature to be extracted are the cracks on the concrete structure. Automating the site inspection will reduce the time taken for the completion of the project and reduce the work load on-site inspectors and make their work safer. The images used for training and validating the Neural Network model are collected from the site and internet databases for data science research. After testing images, the proposed model acquired a decent accuracy of 92% and validation accuracy of 96% which can be deployed for real-world applications.