| Autor: |
M. Pedraza, R.I. Munoz, Marcos Zúñiga, Gonzalo Carvajal, Moulay A. Akhloufi, N.A. Castro, J.M. Quinteros, A. González, B.F. Rosales, C.A. Fernandez, D.G. Cardenas, F.J. Rauh, Christopher Nikulin |
| Rok vydání: |
2017 |
| Předmět: |
|
| Zdroj: |
7th Latin American Conference on Networked and Electronic Media (LACNEM 2017). |
| DOI: |
10.1049/ic.2017.0026 |
| Popis: |
The devastating effects of wildland fires is an unsolved worldwide problem, resulting in human losses and the destruction of natural and economical resources. Assisting firefighters in controlling this kind of natural disasters is an important task. Nowadays, technology advances can help to fulfil this complex task. We propose a new convolutional neural network architecture able to detect fires in images, with high accuracy and high performance which enable the operation of the system in real-time. Preliminary results show that the proposed approach, called SFEwAN-SD, outperforms state-of-the-art approaches both in accuracy and processing time. This performance will be useful in the development of our UAV fire monitoring system that can detect and track fires in real-time. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
