AI-POWERED SYSTEM FOR DETECTING OBSTRUCTIONS BLOCKING STAIRWELLS IN HIGH-RISE BUILDINGS

This article reports a work aiming to identify obstructions in high-rise buildings’ stairwells. Objects left in thestairwells of tall buildings pose a serious safety risk, especially in the event of fire or evacuation. However, toconduct manual inspections of stairwell obstructions is extremely time-consuming and inefficient, especially intall buildings. Artificial intelligence techniques including a neural network computer vision model: YOLOv8,were applied in the prototype of a stairwells obstructions detection system. Dual models of objects were trainedand applied for achieving better flexibility in this kind of solutions. Before widely adopting this kind of system,further works including models that can detect a wider variety of object categories and efficient ways of capturingvideo sources are required.