Feature extraction method of stone curtain wall panels based on multi-source data

Abstract. Various and diverse stone curtain walls serve as typical building facades in civil buildings, which were mainly composed of stone panels and supporting frames, have been widely used in various civil buildings around the world. During long-term service life, stone curtain wall might be subjected to earthquakes, strong winds, environmental changes and other effects, resulting in performance degradation. At present, the monitoring and sensing methods of stone curtain wall panel’s safety and function states are always restricted at application level, and some safety hazards and functional degradation are hardly predicted with efficiency and accuracy, probably resulting in damage accidents, economic losses and negative social impacts during service time. In this paper, multi-source data were achieved through multiple monitoring means, characteristic parameters of mechanical and modeling dimensions were extracted, and the safety and functional performance evaluation of stone panels were investigated. Firstly, vibration signal data were collected by installing acceleration sensors on the curtain wall panel and supporting frame, and image data were analyzed by taking photos with high-definition cameras and drones. Then, the natural vibration frequency and crack length of the mechanical dimension, as well as the crack area and stain area of the molding dimension were extracted and analyzed. Finally, based on the monitoring data, the feature parameters in three dimensions were extracted to evaluate the safety and functional performance of stone panels. The safety and functional performance evaluation method based on multi-source data of stone curtain wall panel is more comprehensive and reliable than traditional methods, which can provide an innovation approach to evaluate and ensure the functional and safety performances of stone curtain wall.