A survey of deep learning-based structural audits

Buildings are built to last for a long time and most of the buildings we see today were built years ago. safety of these buildings is very important. so, structural audits are done regularly to check the structure's health. It covers visual inspection of the buildings, non-destructive tests, and analysis reports. Therefore, structural audits are an essential factor in deciding a building's strength. Today most audits are done manually but there are too many inconsistencies in these practices. However, some of the methodologies can be made efficient and accurate using deep learning processes. for example, audits contain visual inspection of buildings to identify the damage and its severity and suggest repairs to be done. But as this task relies so much on the knowledge and experience of the surveyor it is prone to human errors. Using Deep Learning algorithms to identify and classify the damage based on its severity, will be more accurate and faster than manual inspection. This survey paper tries to summarize current research done in this field, and how we can use these solutions in Structural Audits and overcome current problems using Deep Learning methods such as CNN and intends to generate a more precise framework for these audits.