상세정보
학술대회
Deep Neural Networks for Wild fire Detection with Unmanned Aerial Vehicle
- 발행일
- 201701
- 출처
- International Conference on Consumer Electronics (ICCE) 2017, pp.271-272
- 협약과제
- 16MH4800, 무인기 탑재 복합형 센서 기반의 국지적 재난 감시 및 상황 대응을 위한 스마트 아이 기술 개발, 이용태
- 초록
- Wildfire threatens people's lives and livelihoods. Wildfires kill hundreds of thousand people worldwide each year. Disaster information services are required to save lives and reduce economic losses when wildfire occurs. However, manned airplanes are too expensive to operate for frequent wildland monitoring. Satellite images cannot be used for early wildfire detection due to low temporal resolution and low spatial resolution. Unmanned aerial vehicles are cost-effective means to provide high resolution images for wildfire detection. A wildfire detection system utilizing unmanned aerial vehicles was developed with deep convolutional neural networks. The system achieved high accuracy for wide range of aerial photographs.
- KSP 제안 키워드
- Aerial photographs, Convolution neural network(CNN), Deep convolutional neural networks, Deep neural network(DNN), Disaster information, Economic losses, Fire detection, High accuracy, Information services, Intrusion detection system(IDS), Temporal resolution