Loading [a11y]/accessibility-menu.js
Roof Damage Assessment using Deep Learning | IEEE Conference Publication | IEEE Xplore

Roof Damage Assessment using Deep Learning

Publisher: IEEE

Abstract:

Industrial procedures can be inefficient in terms of time, money and consumer satisfaction. the rivalry among businesses' gradually encourages them to exploit intelligent...View more

Abstract:

Industrial procedures can be inefficient in terms of time, money and consumer satisfaction. the rivalry among businesses' gradually encourages them to exploit intelligent systems to achieve such goals as increasing profits, market share, and higher productivity. The property casualty insurance industry is not an exception. The inspection of a roof's condition is a preliminary stage of the damage claim processing performed by insurance adjusters. When insurance adjusters inspect a roof, it is a time consuming and potentially dangerous endeavor. In this paper, we propose to automate this assessment using RGB imagery of rooftops that have been inflicted with damage from hail impact collected using small unmanned aircraft systems (sUAS) along with deep learning to infer the extent of roof damage (see Fig. I). We assess multiple convolutional neural networks on our unique rooftop damage dataset that was gathered using a sUAS. Our experiments show that we can accurately identify hail damage automatically using our techniques.
Date of Conference: 10-12 October 2017
Date Added to IEEE Xplore: 11 September 2018
ISBN Information:
Electronic ISSN: 2332-5615
Publisher: IEEE
Conference Location: Washington, DC, USA

References

References is not available for this document.