Auto Detection of Roofing Material by Deep Learning Using UAV Photo

To reduce delays in roof repair work after natural disasters caused by typhoons or earthquakes, automatic classification of roofing materials by remote sensing is one of the key technologies for estimating roof repair demand. This paper investigates a technology for classifying roofing materials usi...

Full description

Saved in:
Bibliographic Details
Published inStructural Safety and Reliability: Proceedings of the Japan Conference on Structural Safety and Reliability (JCOSSAR) pp. 167 - 172
Main Authors YOKOYAMA, Hiroto, XU, Junglin, NISHIJIMA, Kazuyoshi, TOMOKIYO, Eriko, TAKEUCHI, Takashi, TAKAHASHI, Toru
Format Journal Article
LanguageJapanese
Published Steering Committee on Japan Conference on Structural Safety and Reliability 2023
Subjects
Classification
Deep learning
Drone
Ortho image
Resolution
Online AccessGet full text

Cover

More Information
Summary:To reduce delays in roof repair work after natural disasters caused by typhoons or earthquakes, automatic classification of roofing materials by remote sensing is one of the key technologies for estimating roof repair demand. This paper investigates a technology for classifying roofing materials using ortho image obtained based on UAV photos. An initial investigation shows that the detection accuracy varies with the resolution of the images to be trained. Therefore, two trained models with different resolutions are used under the same conditions to validate the accuracy. As a result, it is found that detection accuracy is lower when automatic detection is performed using images with a lower resolution than that used in training.
Bibliography:OS9-10A
ISSN:2759-0909
DOI:10.60316/jcossar.10.0_167