Deepcrack: A Deep Learning Approach for Image-Based Crack Prediction using MobileNet And Transfer Learning

Cracks in infrastructure pose significant challenges to public safety and require timely detection for effective maintenance. This research study presents Deep Crack, a novel deep-learning approach for image-based crack prediction. Leveraging the power of Convolutional Neural Networks (CNNs), the Mobile Net architecture is employed as the backbone and implement transfer learning to enhance the model's ability to recognize cracks. The proposed methodology involves extensive data pre-processing, utilizing image augmentation techniques to address data scarcity. The model is trained and validated on a diverse dataset, demonstrating its effectiveness in distinguishing between images containing cracks and those without. The classifier is fine-tuned using a custom top layer, incorporating global average pooling and dense layers. This research work introduces a comprehensive evaluation framework, including confusion matrices and classification reports, to assess the model's performance. Deep Crack achieves remarkable accuracy, precision, and recall in predicting crack presence, demonstrating its capability for practical use in infrastructure maintenance and safety scenarios. This research contributes to the intersection of deep learning and image processing, offering an innovative solution for crack detection. The proposed methodology not only demonstrates the efficacy of Mobile Net and transfer learning but also provides insights into the broader applications of deep learning in civil engineering and infrastructure management.