A Deep Learning Based Object Identification System for Forest Fire Detection
Forest fires are still a large concern in several countries due to the social, environmental and economic damages caused. This paper aims to show the design and validation of a proposed system for the classification of smoke columns with object detection and a deep learning-based approach. This appr...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-10-01
|
| Series: | Fire |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2571-6255/4/4/75 |
| _version_ | 1827672537408995328 |
|---|---|
| author | Federico Guede-Fernández Leonardo Martins Rui Valente de Almeida Hugo Gamboa Pedro Vieira |
| author_facet | Federico Guede-Fernández Leonardo Martins Rui Valente de Almeida Hugo Gamboa Pedro Vieira |
| author_sort | Federico Guede-Fernández |
| collection | DOAJ |
| description | Forest fires are still a large concern in several countries due to the social, environmental and economic damages caused. This paper aims to show the design and validation of a proposed system for the classification of smoke columns with object detection and a deep learning-based approach. This approach is able to detect smoke columns visible below or above the horizon. During the dataset labelling, the smoke object was divided into three different classes, depending on its distance to the horizon, a cloud object was also added, along with images without annotations. A comparison between the use of RetinaNet and Faster R-CNN was also performed. Using an independent test set, an F1-score around 80%, a G-mean around 80% and a detection rate around 90% were achieved by the two best models: both were trained with the dataset labelled with three different smoke classes and with augmentation; Faster R-CNNN was the model architecture, re-trained during the same iterations but following different learning rate schedules. Finally, these models were tested in 24 smoke sequences of the public HPWREN dataset, with 6.3 min as the average time elapsed from the start of the fire compared to the first detection of a smoke column. |
| first_indexed | 2024-03-10T04:09:44Z |
| format | Article |
| id | doaj.art-f8cd12c0eea04ee6947ea88ae9d89a69 |
| institution | Directory Open Access Journal |
| issn | 2571-6255 |
| language | English |
| last_indexed | 2024-03-10T04:09:44Z |
| publishDate | 2021-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Fire |
| spelling | doaj.art-f8cd12c0eea04ee6947ea88ae9d89a692023-11-23T08:14:23ZengMDPI AGFire2571-62552021-10-01447510.3390/fire4040075A Deep Learning Based Object Identification System for Forest Fire DetectionFederico Guede-Fernández0Leonardo Martins1Rui Valente de Almeida2Hugo Gamboa3Pedro Vieira4Physics Department, NOVA School of Science and Technology (Campus de Caparica), 2829-516 Caparica, PortugalPhysics Department, NOVA School of Science and Technology (Campus de Caparica), 2829-516 Caparica, PortugalPhysics Department, NOVA School of Science and Technology (Campus de Caparica), 2829-516 Caparica, PortugalPhysics Department, NOVA School of Science and Technology (Campus de Caparica), 2829-516 Caparica, PortugalPhysics Department, NOVA School of Science and Technology (Campus de Caparica), 2829-516 Caparica, PortugalForest fires are still a large concern in several countries due to the social, environmental and economic damages caused. This paper aims to show the design and validation of a proposed system for the classification of smoke columns with object detection and a deep learning-based approach. This approach is able to detect smoke columns visible below or above the horizon. During the dataset labelling, the smoke object was divided into three different classes, depending on its distance to the horizon, a cloud object was also added, along with images without annotations. A comparison between the use of RetinaNet and Faster R-CNN was also performed. Using an independent test set, an F1-score around 80%, a G-mean around 80% and a detection rate around 90% were achieved by the two best models: both were trained with the dataset labelled with three different smoke classes and with augmentation; Faster R-CNNN was the model architecture, re-trained during the same iterations but following different learning rate schedules. Finally, these models were tested in 24 smoke sequences of the public HPWREN dataset, with 6.3 min as the average time elapsed from the start of the fire compared to the first detection of a smoke column.https://www.mdpi.com/2571-6255/4/4/75smoke detectionfire detectionwildfiresdeep learning |
| spellingShingle | Federico Guede-Fernández Leonardo Martins Rui Valente de Almeida Hugo Gamboa Pedro Vieira A Deep Learning Based Object Identification System for Forest Fire Detection Fire smoke detection fire detection wildfires deep learning |
| title | A Deep Learning Based Object Identification System for Forest Fire Detection |
| title_full | A Deep Learning Based Object Identification System for Forest Fire Detection |
| title_fullStr | A Deep Learning Based Object Identification System for Forest Fire Detection |
| title_full_unstemmed | A Deep Learning Based Object Identification System for Forest Fire Detection |
| title_short | A Deep Learning Based Object Identification System for Forest Fire Detection |
| title_sort | deep learning based object identification system for forest fire detection |
| topic | smoke detection fire detection wildfires deep learning |
| url | https://www.mdpi.com/2571-6255/4/4/75 |
| work_keys_str_mv | AT federicoguedefernandez adeeplearningbasedobjectidentificationsystemforforestfiredetection AT leonardomartins adeeplearningbasedobjectidentificationsystemforforestfiredetection AT ruivalentedealmeida adeeplearningbasedobjectidentificationsystemforforestfiredetection AT hugogamboa adeeplearningbasedobjectidentificationsystemforforestfiredetection AT pedrovieira adeeplearningbasedobjectidentificationsystemforforestfiredetection AT federicoguedefernandez deeplearningbasedobjectidentificationsystemforforestfiredetection AT leonardomartins deeplearningbasedobjectidentificationsystemforforestfiredetection AT ruivalentedealmeida deeplearningbasedobjectidentificationsystemforforestfiredetection AT hugogamboa deeplearningbasedobjectidentificationsystemforforestfiredetection AT pedrovieira deeplearningbasedobjectidentificationsystemforforestfiredetection |