ABSTRACT
“Monuments are the grappling manacles that bind one generation to another”—quoted by French Poet Joseph Jubert. Historical buildings are the pride of all nations and bear good signs of their history and culture. Their protection deserves great attention because it is extremely important from historical, cultural, and economic perspectives. The preservation of historical structures plays a key role in preserving our culture and heritage. Finding and assessing surface damage in historic buildings using vision-based manual inspection methods require much time and effort. Currently, the basis of the method of detecting surface damage of historical cultural objects is on-site visual inspection methods confirmed by specialized equipment. This approach has significant disadvantages; inspectors without expertise can misjudge damage, which can have a significant negative impact on subsequent structural safety assessment and repair. This method of conducting large-scale research on the building is time-consuming and inefficient. In addition, the subsequent analysis and processing of the data require numerous qualified workers, which require time and a lot of work. Therefore, this method is not sufficient for the rapid detection of damage in historical buildings. This chapter proposes a unique unmanned aerial vehicle (UAV) image acquisition system and automatic damage detection method using mask region-based convolutional neural network (Mask R-CNN) enveloped with a mask model based on the DenseNet structure.
