Site Loader
Rengerslaan 8-10, 8971DD Leeuwarden

We propose an automated classification approach using CNNs to detect and classify
building components. Aerial images are analyzed, focusing on window casings and
roof tiles. Models like VGG11 and ResNet50 demonstrate outstanding performance,
showcasing the potential of deep learning in defect identification. This research
contributes to automating building inspections, providing an efficient solution for
real-world scenarios

Post Author: Koen Molenaar

Leave a Reply