Enhancing Deep Learning Model Performance for Surface Crack Detection

This research investigates the application of deep learning models for early detection of surface cracks in infrastructure. By automating this process, the study aims to improve structural maintenance and ensure public safety. This research study compares five pre-trained deep learning models: InceptionV3, VGG19, ResNet50, Xception, and MobileNet. The results demonstrate the effectiveness of deep learning in accurately detecting surface cracks, even in their early stages. By utilizing advanced deep learning techniques, this research provides a valuable tool for stakeholders to monitor the health of infrastructure and identify potential risks. Early detection enables timely maintenance, contributing to a safer and more durable built environment. This study highlights the importance of incorporating deep learning into infrastructure safety practices to enhance structural integrity and protect public safety.