Automated classification of fuel types using roadside images via deep learning
Md Riasat Azim
A , Melih Keskin B , Ngoan Do A and Mustafa Gül A *
+ Author Affiliations
- Author Affiliations
A Civil and Environmental Engineering, University of Alberta, Edmonton, Canada.
B Computing Science, University of Alberta, Edmonton, Canada.
* Correspondence to: mustafa.gul@ualberta.ca
International Journal of Wildland Fire 31(10) 982-987 https://doi.org/10.1071/WF21136
Submitted: 30 September 2021 Accepted: 9 August 2022 Published: 2 September 2022
© 2022 The Author(s) (or their employer(s)). Published by CSIRO Publishing on behalf of IAWF.
Abstract
There is an urgent need to develop new data-driven methods for assessing wildfire-related risks in large areas susceptible to such risks. To assess these risks, one of the critical parameters to analyse is fuel. This research note presents a framework for classifying fuels through the analysis of roadside images to complement the current practice of assessing fuels through aerial images and visual inspections. Some of the most prevalent fuel types in North America were considered for automated classification, including grasses, shrubs and timbers. A framework was developed using convolutional neural networks (CNNs), which can automate the process of fuel classification. Various pre-trained neural networks were examined and the best network in terms of time efficiency and accuracy was identified, and had ~94% accuracy in identifying the chosen fuel types. This framework was initially applied to street view images collected from Google Earth. Indeed, the results showed that the framework has the potential for application for fuel classification using roadside images, and this makes it suitable for crowdsensing-based fuel mapping for wildfire risk assessment, which is the future goal of this research.
Keywords: automated fuel identification framework, convolutional neural network, deep learning, fuel classification, fuel identification, North American fuels, pre-trained networks, road-side image analysis.
References
Anderson HE (1982) Aids to determining fuel models for estimating fire behavior. General Technical Report INT-122. (USDA Forest Service, Intermountain Forest, and Range Experiment Station: Ogden, Utah)| Crossref |
Ayyachamy S, Alex V, Khened M, Krishnamurthi G (2019) Medical image retrieval using Resnet-18. In ‘Proceedings SPIE 10954, Medical Imaging 2019: Imaging Informatics for Healthcare, Research, and Applications’. p. 1095410. (SPIE)
| Crossref |
Bénichou N, Adelzadeh M, Singh J, Gomaa I, Elsagan N, Kinateder M, Ma C, Gaur A, Bwalya A, Sultan M (2021) ‘National guide for wildland-urban interface fires.’ (National Research Council Canada: Ottawa, ON)
Benito A, Hernandez-Leal PA, Arbelo M, Gonzalez-Calvo A, Moreno-Ruiz JA, Garcia-Lazaro JR (2016) Satellite image based methods for fuels maps updating. In ‘Proceedings SPIE 9998, Remote Sensing for Agriculture, Ecosystems, and Hydrology XVIII’. p. 999821. (SPIE)
| Crossref |
Cencerrado A, Cortes A, Margalef T (2014) Response time assessment in forest fire spread simulation: An integrated methodology for efficient exploitation of available prediction time. Environmental Modelling & Software 54, 153–164.
| Response time assessment in forest fire spread simulation: An integrated methodology for efficient exploitation of available prediction time.Crossref | GoogleScholarGoogle Scholar |
Chirici G, Scotti R, Montaghi A, Barbati A, Cartisano R, Lopez G, Marchetti M, Mcroberts RE, Olsson H, Corona P (2013) Stochastic gradient boosting classification trees for forest fuel types mapping through airborne laser scanning and IRS LISS-III imagery. International Journal of Applied Earth Observation and Geoinformation 25, 87–97.
| Stochastic gradient boosting classification trees for forest fuel types mapping through airborne laser scanning and IRS LISS-III imagery.Crossref | GoogleScholarGoogle Scholar |
Denham M, Laneri K (2018) Using efficient parallelization in Graphic Processing Units to parameterize stochastic fire propagation models. Journal of Computational Science 25, 76–88.
| Using efficient parallelization in Graphic Processing Units to parameterize stochastic fire propagation models.Crossref | GoogleScholarGoogle Scholar |
Feng C, Zhang H, Wang S, Li Y, Wang H, Yan F (2019) Structural damage detection using deep convolutional neural network and transfer learning. KSCE Journal of Civil Engineering 23, 4493–4502.
| Structural damage detection using deep convolutional neural network and transfer learning.Crossref | GoogleScholarGoogle Scholar |
García M, Riaño D, Chuvieco E, Salas J, Danson FM (2011) Multispectral and LiDAR data fusion for fuel type mapping using Support Vector Machine and decision rules. Remote Sensing of Environment 115, 1369–1379.
| Multispectral and LiDAR data fusion for fuel type mapping using Support Vector Machine and decision rules.Crossref | GoogleScholarGoogle Scholar |
Guo M, Du Y (2019) Classification of thyroid ultrasound standard plane images using ResNet-18 Networks. In ‘2019 IEEE 13th International Conference on Anti-counterfeiting, Security, and Identification (ASID)’. pp. 324–328. (IEEE)
| Crossref |
Jain P, Coogan SCP, Subramanian SG, Crowley M, Taylor S, Flannigan MD (2020) A review of machine learning applications in wildfire science and management. Environmental Reviews 28, 478–505.
| A review of machine learning applications in wildfire science and management.Crossref | GoogleScholarGoogle Scholar |
Jing E, Zhang H, Li Z, Liu Y, Ji Z, Ganchev I (2021) ECG heartbeat classification based on an improved ResNet-18 model. Computational and Mathematical Methods in Medicine 2021, 6649970
| ECG heartbeat classification based on an improved ResNet-18 model.Crossref | GoogleScholarGoogle Scholar |
Joshi R (2016) Accuracy, precision, recall and F1 score: Interpretation of performance measures. Available at https://blog.exsilio.com/all/accuracy-precision-recall-f1-score-interpretation-of-performance-measures/
Kochgaven C, Mishra P, Shitole S (2021) Detecting Presence of COVID-19 with ResNet-18 using PyTorch. In ‘2021 International Conference on Communication information and Computing Technology (ICCICT)’. pp. 1–6. (IEEE)
| Crossref |
Lachowski H, Maus P, Golden M, Johnson J, Landrum V, Powell J, Varner V, Wirth T, Gonzales J, Bain S (1995) Guidelines for the use of digital imagery for vegetation mapping. USDA Forest Service Engineering Staff EM-7140-25; Washington DC.
LANDFIRE (2021) Conterminous US LANDFIRE LF 2020 (LF_220) Vegetation us_220 Existing Vegetation Type Map. (USDA Forest Service and US Deparment of the Interior). Available at https://landfire.cr.usgs.gov/viewer/viewer.html
Lanorte A, Lasaponara R (2011) On the use of satellite remote sensing data to characterize and map fuel types. In ‘Computational Science, and Its Applications – ICCSA 2011, Lecture Notes in Computer Science’. (Eds B Murgante, O Gervasi, A Iglesias, D Taniar, BO Apduhan) Vol. 6783. (Springer: Berlin, Heidelberg)
| Crossref |
López-Serrano PM, López-Sánchez CA, Álvarez-González JG, García-Gutiérrez J (2016) A comparison of machine learning techniques applied to Landsat-5 TM spectral data for biomass estimation. Canadian Journal of Remote Sensing 42, 690–705.
| A comparison of machine learning techniques applied to Landsat-5 TM spectral data for biomass estimation.Crossref | GoogleScholarGoogle Scholar |
Makantasis K, Protopapadakis E, Doulamis A, Doulamis N, Loupos C, Nikolaos D (2015) Deep convolutional neural networks for efficient vision based tunnel inspection. In ‘IEEE International Conference on Intelligent Computer Communication and Processing 2015’. pp. 335–342. (Romania ICCP: Cluj-Napoca)
MATLAB (2021) ‘version R2021a.’ (The MathWorks Inc: Natick, Massachusetts)
McGregor S, Houtman R, Montgomery C, Metoyer R, Dietterich GT (2017) Fast optimization of wildfire suppression policies with SMAC. arXiv preprint Available at http://arxiv.org/abs/1703.09391
Mei Q, Gül M, Azim MR (2020) Densely connected deep neural network considering connectivity of pixels for automatic crack detection. Automation in Construction 110, 103018
| Densely connected deep neural network considering connectivity of pixels for automatic crack detection.Crossref | GoogleScholarGoogle Scholar |
Oswald BP, Fancher JT, Kulhavy DL, Reeves HC (1999) Classifying fuels with aerial photography in east Texas. International Journal of Wildland Fire 9, 109–113.
| Classifying fuels with aerial photography in east Texas.Crossref | GoogleScholarGoogle Scholar |
Penman TD, Price O, Bradstock RA (2011) Bayes Nets as a method for analysing the influence of management actions in fire planning. International Journal of Wildland Fire 20, 909–920.
| Bayes Nets as a method for analysing the influence of management actions in fire planning.Crossref | GoogleScholarGoogle Scholar |
Pierce AD, Farris CA, Taylor AH (2012) Use of random forests for modeling and mapping forest canopy fuels for fire behavior analysis in Lassen Volcanic National Park, California, USA Forest Ecology and Management 279, 77–89.
| Use of random forests for modeling and mapping forest canopy fuels for fire behavior analysis in Lassen Volcanic National Park, California, USACrossref | GoogleScholarGoogle Scholar |
Riano D, Ustin DS, Usero L, Patricio M (2005) ‘Estimation of fuel moisture content using neural networks’. pp. 489–498. (Springer: Berlin, Heidelberg)
| Crossref |
Riley KL, Grenfell IC, Finney MA, Crookston NL (2014) Utilizing random forests imputation of forest plot data for landscape-level wildfire analyses. In ‘Advances in Forest Fire Research’. pp. 583–590. (Ed. DX Viegas) (Imprensa da Universidade de Coimbra: Coimbra).
| Crossref |
Sayad YO, Mousannif H, Al Moatassime H (2019) Predictive modeling of wildfires: A new dataset and machine learning approach. Fire Safety Journal 104, 130–146.
| Predictive modeling of wildfires: A new dataset and machine learning approach.Crossref | GoogleScholarGoogle Scholar |
Shorten C, Khoshgoftaar TM (2019) A survey on image data augmentation for deep learning. Journal of Big Data 6, 60
| A survey on image data augmentation for deep learning.Crossref | GoogleScholarGoogle Scholar |
Soukup D, Huber-Mörk R (2014) Convolutional neural networks for steel surface defect detection from photometric stereo images. In ‘Advances in Visual Computing. ISVC 2014. Lecture Notes in Computer Science’. (Eds Bebis G, et al.) pp. 668–677. (Springer: Cham)
| Crossref |
Zhao J, Zhang Z, Han S, Qu C, Yuan Z, Zhang D (2011) SVM based forest fire detection using static and dynamic features. Computer Science and Information Systems 8, 821–841.
| SVM based forest fire detection using static and dynamic features.Crossref | GoogleScholarGoogle Scholar |
Zhu J, Song J (2020) An intelligent classification model for surface defects on cement concrete bridges. Applied Sciences 10, 972–990.
| An intelligent classification model for surface defects on cement concrete bridges.Crossref | GoogleScholarGoogle Scholar |
