Crack Detection in Concrete Structures Using Image Processing and Deep Learning

摘要

This paper proposes a combined approach involving image processing and deep learning algorithms to detect the cracks present in building structures such as bridges and other massive structures. The deep learning model used in the proposed work is the Mask R-CNN model. The model is trained on a total of 40,000 images with the help of a Supervisely software to perform crack detection. To distinguish the results, the images are also segmented using active contour model and Chan-Vese segmentation algorithm. The segmented images obtained are trained and validated using a fully convolutional network model. The results obtained on using the pre-trained model and FCNN algorithms are detailed and the study yields effective research. The study shows that deep learning scheme results in an alternative to the current visual examination.