Automatic assessment of roofs conditions using artificial intelligence (AI) and unmanned aerial vehicles (UAVs)

Abstract

Building roof inspections must be performed regularly to ensure repairs and replacements are done promptly. These inspections get overlooked on sloped roofs due to the inefficiency of manual inspections and the difficulty of accessing sloped roofs. Walking a roof to inspect each tile is time-consuming, and as the roof slope increases, this difficulty increases the time needed for an inspection. Moreover, there is an intrinsic safety risk involved. Falls from roofs tend to cause severe and expensive injuries. The emergence of new sensing technologies and artificial intelligence (AI) such as high-resolution imagery and deep learning has enabled humans to move beyond the concept of using manual labor in damage assessments. It has brought significant advantages in the field of safety management, and it can be a substitute for the traditional assessment of roofs. This study uses unmanned aerial vehicles (UAVs) and deep learning technology to perform sloped roof inspections effectively, thus eliminating the safety risk involved in traditional manual inspections. This study utilizes UAVs and deep learning to automatically collect and classify roof imagery to identify missing shingles on the roof. The proposed research can help real estate agents, insurance companies, and others make better and more informed decisions about roof conditions. Future research could be refining the model to deal with different types of defects in addition to missing shingles.

A Transformer-Optimized Deep Learning Network for Road Damage Detection and Tracking

“…Ahmet Bahaddin Ersoz et al [39] processed a UAV-based pavement crack recognition system by processing UAV-based images for support vector machine (SVM) model training. Ammar Alzarrad et al [40] demonstrated the effectiveness of combining AI and UAVs by combining high-resolution imagery with deep learning to detect disease on roofs. Long Ngo Hoang, T et al [41] presented a methodology based on the mask regions with a convolutional neural network model, which was coupled with the new object detection framework Detectron2 to train a model that utilizes roadway imagery acquired from an unmanned aerial system (UAS).…”

Eyes in the Sky: Drones Applications in the Built Environment under Climate Change Challenges

“…Utilizing infrared sensors also highlights areas of moisture intrusion within wall systems signaled by temperature differentials [119]. A study by Alzarrad et al [159] presented a promising approach to automating sloped roof inspections using UAVs and deep learning. It highlighted the significance of such technology in enhancing efficiency, reducing safety risks, and providing accurate assessments of roof conditions.…”