Crack Detection of Old Residential Buildings Based on UAV Intelligent Vision

Old residential areas have long been a weak link in urban management, and building surface defects frequently occur, seriously endangering social and public safety. As a typical representative of building surface defects, building surface cracks are one of the main factors affecting building safety. This paper proposes a crack detection method for old residential buildings based on unmanned aerial vehicle (UAV) intelligent vision, which mainly includes image preprocessing and surface defect image recognition. First, median filter preprocessing is performed on the image data with and without cracks collected by the UAV. Next, to improve the robustness of the model, data augmentation is performed on the preprocessed image data. Finally, a Convolutional Neural Network (CNN) is established to realize the surface recognition of defect images. Confusion matrix evaluation and confidence test results demonstrate the feasibility and effectiveness of the proposed method.