Automatic detection of foreign objects on building roof surfaces

The continuously advancing and ever-evolving technology has also provided significant momentum in the field of photogrammetry. Unmanned Aerial Vehicles, through the data they produce within predefined flight plans, serve as valuable sources for numerous disciplines. Foreign objects that accumulate over time or emerge due to various factors on rooftop surfaces can lead to serious issues in terms of both structural safety and maintenance processes. Therefore, the rapid and reliable detection of such objects has become a critical requirement for monitoring building health and mitigating risks in urban areas. This study, which focuses on detecting foreign objects on building rooftops, was conducted using UAV imagery within the MATLAB environment. The reliability of the obtained results was evaluated through accuracy analysis. The accuracy rate was calculated as 88.66%, completeness rate as 100%, and the F1-score as 93.33%. These findings indicate that the proposed approach can be effectively utilized, particularly in studies based on change detection and object analysis. Furthermore, the developed method has the potential to be applied across various domains, including urban monitoring, post-disaster damage assessment, sustainable environmental planning, and maintenance scheduling.