Building Surface Crack Detection Using Deep Learning Technology

Cracks in building facades are inevitable due to the age of building. found facade may be further exacerbated if not corrected immediately. Considering extensive size some buildings, there is definitely a need automate inspection routine facilitate process. The incorporation deep learning technology for classification images has proven an effective method many past civil infrastructures like pavements and bridges. There is, however, limited research built environment sector. In order align with Smart Nation goals country, use technologies necessary construction industry. focus study identify effectiveness image classification. Deep technology, such as Convolutional Neural Networks (CNN), requires large amount data obtain good performance. It difficult collect manually. This will cover transfer approach, where can carried out even data. Using CNN achieved accuracy level about 89%, while using model 94%. Based on this, it concluded that achieves better performance compared same input.