{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,9]],"date-time":"2026-03-09T21:44:57Z","timestamp":1773092697218,"version":"3.50.1"},"reference-count":51,"publisher":"MDPI AG","issue":"11","license":[{"start":{"date-parts":[[2023,5,24]],"date-time":"2023-05-24T00:00:00Z","timestamp":1684886400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"NASA ROSES AIST-QRS-21","award":["NNA16BD14C"],"award-info":[{"award-number":["NNA16BD14C"]}]},{"name":"NASA Academic Mission Services","award":["NNA16BD14C"],"award-info":[{"award-number":["NNA16BD14C"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Remote Sensing"],"abstract":"<jats:p>Wildfires are one of the major disasters among many and are responsible for more than 6 million acres burned in the United States alone every year. Accurate, insightful, and timely wildfire detection is needed to help authorities mitigate and prevent further destruction. Uncertainty quantification is always a crucial part of the detection of natural disasters, such as wildfires, and modeling products can be misinterpreted without proper uncertainty quantification. In this study, we propose a supervised deep generative machine-learning model that generates stochastic wildfire detection, allowing fast and comprehensive uncertainty quantification for individual and collective events. In the proposed approach, we also aim to address the patchy and discontinuous Moderate Resolution Imaging Spectroradiometer (MODIS) wildfire product by training the proposed model with MODIS raw and combined bands to detect fire. This approach allows us to generate diverse but plausible segmentations to represent the disagreements regarding the delineation of wildfire boundaries by subject matter experts. The proposed approach generates stochastic segmentation via two model streams in which one learns meaningful stochastic latent distributions, and the other learns the visual features. Two model branches join eventually to become a supervised stochastic image-to-image wildfire detection model. The model is compared to two baseline stochastic machine-learning models: (1) with permanent dropout in training and test phases and (2) with Stochastic ReLU activations. The visual and statistical metrics demonstrate better agreements between the ground truth and the proposed model segmentations. Furthermore, we used multiple scenarios to evaluate the model comprehension, and the proposed Probabilistic U-Net model demonstrates a better understanding of the underlying physical dynamics of wildfires compared to the baselines.<\/jats:p>","DOI":"10.3390\/rs15112718","type":"journal-article","created":{"date-parts":[[2023,5,24]],"date-time":"2023-05-24T01:59:13Z","timestamp":1684893553000},"page":"2718","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":12,"title":["Probabilistic Wildfire Segmentation Using Supervised Deep Generative Model from Satellite Imagery"],"prefix":"10.3390","volume":"15","author":[{"given":"Ata","family":"Akbari Asanjan","sequence":"first","affiliation":[{"name":"Data Science Group (DSG), NASA Ames Research Center, Moffett Field, CA 94035, USA"},{"name":"USRA Research Institute for Advanced Computer Science (RIACS), Washington, DC 20024, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7672-6033","authenticated-orcid":false,"given":"Milad","family":"Memarzadeh","sequence":"additional","affiliation":[{"name":"Data Science Group (DSG), NASA Ames Research Center, Moffett Field, CA 94035, USA"},{"name":"USRA Research Institute for Advanced Computer Science (RIACS), Washington, DC 20024, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9718-9641","authenticated-orcid":false,"given":"Paul Aaron","family":"Lott","sequence":"additional","affiliation":[{"name":"USRA Research Institute for Advanced Computer Science (RIACS), Washington, DC 20024, USA"},{"name":"Quantum Artificial Intelligence Laboratory (QuAIL), NASA Ames Research Center, Moffett Field, CA 94035, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Eleanor","family":"Rieffel","sequence":"additional","affiliation":[{"name":"Quantum Artificial Intelligence Laboratory (QuAIL), NASA Ames Research Center, Moffett Field, CA 94035, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shon","family":"Grabbe","sequence":"additional","affiliation":[{"name":"Quantum Artificial Intelligence Laboratory (QuAIL), NASA Ames Research Center, Moffett Field, CA 94035, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2023,5,24]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"350","DOI":"10.1139\/x90-051","article-title":"Forest fire history of desolation peak, Washington","volume":"20","author":"Agee","year":"1990","journal-title":"Can. J. For. Res."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"e2009717118","DOI":"10.1073\/pnas.2009717118","article-title":"Warming enabled upslope advance in western US forest fires","volume":"118","author":"Alizadeh","year":"2021","journal-title":"Proc. Natl. Acad. Sci. USA"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"e02161","DOI":"10.1002\/ecs2.2161","article-title":"Wildfires managed for restoration enhance ecological resilience","volume":"9","author":"Barros","year":"2018","journal-title":"Ecosphere"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s40663-015-0033-8","article-title":"Negative consequences of positive feedbacks in US wildfire management","volume":"2","author":"Calkin","year":"2015","journal-title":"For. Ecosyst."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"164","DOI":"10.1016\/j.apgeog.2014.11.025","article-title":"Human-ignited wildfire patterns and responses to policy shifts","volume":"56","author":"Prestemon","year":"2015","journal-title":"Appl. Geogr."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"2928","DOI":"10.1002\/2014GL059576","article-title":"Large wildfire trends in the western United States, 1984\u20132011","volume":"41","author":"Dennison","year":"2014","journal-title":"Geophys. Res. Lett."},{"key":"ref_7","unstructured":"Hoover, K., and Hanson, L.A. (2021). Wildfire Statistics, Congressional Research Service. Technical Report."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"72","DOI":"10.1016\/j.rse.2018.08.005","article-title":"The Collection 6 MODIS burned area mapping algorithm and product","volume":"217","author":"Giglio","year":"2018","journal-title":"Remote Sens. Environ."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"478","DOI":"10.1139\/er-2020-0019","article-title":"A review of machine learning applications in wildfire science and management","volume":"28","author":"Jain","year":"2020","journal-title":"Environ. Rev."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"169","DOI":"10.1016\/j.landurbplan.2010.11.017","article-title":"Land use and topography influences on wildfire occurrence in northern Portugal","volume":"100","author":"Carmo","year":"2011","journal-title":"Landsc. Urban Plan."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"878","DOI":"10.1016\/j.apgeog.2011.09.004","article-title":"Influences of forest roads on the spatial patterns of human-and lightning-caused wildfire ignitions","volume":"32","author":"Narayanaraj","year":"2012","journal-title":"Appl. Geogr."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Ghali, R., Akhloufi, M.A., Jmal, M., Souidene Mseddi, W., and Attia, R. (2021). Wildfire segmentation using deep vision transformers. Remote Sens., 13.","DOI":"10.3390\/rs13173527"},{"key":"ref_13","unstructured":"Green, M.E. (2020). Some Results on a Set of Data Driven Stochastic Wildfire Models, The University of Vermont and State Agricultural College."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Khryashchev, V., and Larionov, R. (2020, January 11\u201313). Wildfire segmentation on satellite images using deep learning. Proceedings of the 2020 Moscow Workshop on Electronic and Networking Technologies (MWENT), Moscow, Russia.","DOI":"10.1109\/MWENT47943.2020.9067475"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"7001","DOI":"10.1109\/JSTARS.2021.3093625","article-title":"Wildfire detection from multisensor satellite imagery using deep semantic segmentation","volume":"14","author":"Rashkovetsky","year":"2021","journal-title":"IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"1244","DOI":"10.1080\/01621459.2000.10474324","article-title":"Inference for deterministic simulation models: The Bayesian melding approach","volume":"95","author":"Poole","year":"2000","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"153","DOI":"10.1016\/j.envsoft.2014.01.008","article-title":"Response time assessment in forest fire spread simulation: An integrated methodology for efficient exploitation of available prediction time","volume":"54","author":"Cencerrado","year":"2014","journal-title":"Environ. Model. Softw."},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Toan, N.T., Cong, P.T., Hung, N.Q.V., and Jo, J. (2019, January 1\u20133). A deep learning approach for early wildfire detection from hyperspectral satellite images. Proceedings of the 2019 7th International Conference on Robot Intelligence Technology and Applications (RiTA), Daejeon, Republic of Korea.","DOI":"10.1109\/RITAPP.2019.8932740"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"130","DOI":"10.1016\/j.firesaf.2019.01.006","article-title":"Predictive modeling of wildfires: A new dataset and machine learning approach","volume":"104","author":"Sayad","year":"2019","journal-title":"Fire Saf. J."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"47","DOI":"10.2495\/FIVA100051","article-title":"Statistical parameter estimation for a cellular automata wildfire model based on satellite observations","volume":"137","author":"Couce","year":"2010","journal-title":"WIT Trans. Ecol. Environ."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"5","DOI":"10.3389\/fmech.2019.00005","article-title":"Modeling wind direction distributions using a diagnostic model in the context of probabilistic fire spread prediction","volume":"5","author":"Quill","year":"2019","journal-title":"Front. Mech. Eng."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"163","DOI":"10.1146\/annurev.earth.33.092203.122552","article-title":"Representing model uncertainty in weather and climate prediction","volume":"33","author":"Palmer","year":"2005","journal-title":"Annu. Rev. Earth Planet. Sci."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"e3837","DOI":"10.1002\/cpe.3837","article-title":"Time aware genetic algorithm for forest fire propagation prediction: Exploiting multi-core platforms","volume":"29","author":"Cencerrado","year":"2017","journal-title":"Concurr. Comput. Pract. Exp."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1016\/j.jocs.2018.02.007","article-title":"Using efficient parallelization in graphic processing units to parameterize stochastic fire propagation models","volume":"25","author":"Denham","year":"2018","journal-title":"J. Comput. Sci."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"133505","DOI":"10.1016\/j.scitotenv.2019.07.311","article-title":"Stochastic decision trigger modelling to assess the probability of wildland fire impact","volume":"694","author":"Ramirez","year":"2019","journal-title":"Sci. Total Environ."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"105050","DOI":"10.1016\/j.envsoft.2021.105050","article-title":"Multifidelity prediction in wildfire spread simulation: Modeling, uncertainty quantification and sensitivity analysis","volume":"141","author":"Valero","year":"2021","journal-title":"Environ. Model. Softw."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"4545","DOI":"10.1029\/2017JD027823","article-title":"Comparison of fire radiative power estimates from VIIRS and MODIS observations","volume":"123","author":"Li","year":"2018","journal-title":"J. Geophys. Res. Atmos."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Allison, R.S., Johnston, J.M., Craig, G., and Jennings, S. (2016). Airborne optical and thermal remote sensing for wildfire detection and monitoring. Sensors, 16.","DOI":"10.3390\/s16081310"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Schultz, C.J., Nauslar, N.J., Wachter, J.B., Hain, C.R., and Bell, J.R. (2019). Spatial, temporal and electrical characteristics of lightning in reported lightning-initiated wildfire events. Fire, 2.","DOI":"10.3390\/fire2020018"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"1798","DOI":"10.1109\/TPAMI.2013.50","article-title":"Representation learning: A review and new perspectives","volume":"35","author":"Bengio","year":"2013","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Akbari Asanjan, A., Das, K., Li, A., Chirayath, V., Torres-Perez, J., and Sorooshian, S. (2020, January 6\u201310). Learning instrument invariant characteristics for generating high-resolution global coral reef maps. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Virtual Event.","DOI":"10.1145\/3394486.3403312"},{"key":"ref_32","unstructured":"Ronneberger, O., Fischer, P., and Brox, T. (2015). Medical Image Computing and Computer-Assisted Intervention\u2014MICCAI 2015: 18th International Conference, Munich, Germany, 5\u20139 October 2015, Springer International Publishing."},{"key":"ref_33","unstructured":"Shu, R., Brofos, J., Zhang, F., Bui, H.H., Ghavamzadeh, M., and Kochenderfer, M. (2016, January 8\u201310). Stochastic video prediction with conditional density estimation. Proceedings of the ECCV Workshop on Action and Anticipation for Visual Learning, Amsterdam, The Netherlands."},{"key":"ref_34","first-page":"6965","article-title":"A Probabilistic U-Net for Segmentation of Ambiguous Images","volume":"31","author":"Kohl","year":"2018","journal-title":"NeurIPS"},{"key":"ref_35","unstructured":"Kingma, D., and Welling, M. (2013). Auto-encoding variational bayes. arXiv."},{"key":"ref_36","unstructured":"Kingma, D., Rezende, D., Mohamed, S., and Welling, M. (2014). Semi-supervised learning with deep generative models. Adv. Neural Inf. Process. Syst., 27."},{"key":"ref_37","first-page":"3483","article-title":"Learning structured output representation using deep conditional generative models","volume":"28","author":"Sohn","year":"2015","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_38","unstructured":"Frazier-Logue, N., and Hanson, S.J. (2018). Dropout is a special case of the stochastic delta rule: Faster and more accurate deep learning. arXiv."},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"265","DOI":"10.1016\/0167-2789(90)90081-Y","article-title":"A stochastic version of the delta rule","volume":"42","author":"Hanson","year":"1990","journal-title":"Phys. D Nonlinear Phenom."},{"key":"ref_40","unstructured":"Kendall, A., Badrinarayanan, V., and Cipolla, R. (2015). Bayesian segnet: Model uncertainty in deep convolutional encoder-decoder architectures for scene understanding. arXiv."},{"key":"ref_41","unstructured":"Gulcehre, C., Moczulski, M., Denil, M., and Bengio, Y. (2016, January 20\u201322). Noisy activation functions. Proceedings of the International Conference on Machine Learning, New York, NY, USA."},{"key":"ref_42","unstructured":"Shridhar, K., Lee, J., Hayashi, H., Mehta, P., Iwana, B.K., Kang, S., Uchida, S., Ahmed, S., and Dengel, A. (2019). Probact: A probabilistic activation function for deep neural networks. arXiv."},{"key":"ref_43","doi-asserted-by":"crossref","first-page":"85","DOI":"10.1016\/j.rse.2013.12.008","article-title":"The New VIIRS 375 m active fire detection data product: Algorithm description and initial assessment","volume":"143","author":"Schroeder","year":"2014","journal-title":"Remote Sens. Environ."},{"key":"ref_44","doi-asserted-by":"crossref","first-page":"418","DOI":"10.1016\/j.ecoinf.2010.04.001","article-title":"Wildfire potential evaluation during a drought event with a regional climate model and NDVI","volume":"5","author":"Liu","year":"2010","journal-title":"Ecol. Inform."},{"key":"ref_45","doi-asserted-by":"crossref","first-page":"138","DOI":"10.1016\/j.rse.2006.06.023","article-title":"Evaluating remotely sensed live fuel moisture estimations for fire behavior predictions in Georgia, USA","volume":"108","author":"Dasgupta","year":"2007","journal-title":"Remote Sens. Environ."},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"295","DOI":"10.1016\/0034-4257(88)90106-X","article-title":"A soil-adjusted vegetation index (SAVI)","volume":"25","author":"Huete","year":"1988","journal-title":"Remote Sens. Environ."},{"key":"ref_47","doi-asserted-by":"crossref","first-page":"2636","DOI":"10.3390\/s7112636","article-title":"Sensitivity of the enhanced vegetation index (EVI) and normalized difference vegetation index (NDVI) to topographic effects: A case study in high-density cypress forest","volume":"7","author":"Matsushita","year":"2007","journal-title":"Sensors"},{"key":"ref_48","doi-asserted-by":"crossref","unstructured":"Pereira-Pires, J.E., Aubard, V., Ribeiro, R.A., Fonseca, J.M., Silva, J.M., and Mora, A. (2021, January 11\u201316). Fuel Break Vegetation Monitoring with Sentinel-2 NDVI Robust to Phenology and Environmental Conditions. Proceedings of the 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, Brussels, Belgium.","DOI":"10.1109\/IGARSS47720.2021.9554943"},{"key":"ref_49","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/LSENS.2021.3112822","article-title":"Low-Cost NDVI Platform for Land Operation: Passive and Active","volume":"5","author":"Mazeh","year":"2021","journal-title":"IEEE Sens. Lett."},{"key":"ref_50","doi-asserted-by":"crossref","first-page":"822","DOI":"10.1071\/WF20077_CO","article-title":"Corrigendum to: Integrating remotely sensed fuel variables into wildfire danger assessment for China","volume":"30","author":"Quan","year":"2021","journal-title":"Int. J. Wildland Fire"},{"key":"ref_51","doi-asserted-by":"crossref","first-page":"59","DOI":"10.1016\/j.foreco.2016.08.035","article-title":"Weather, fuels, and topography impede wildland fire spread in western US landscapes","volume":"380","author":"Holsinger","year":"2016","journal-title":"For. Ecol. Manag."}],"container-title":["Remote Sensing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2072-4292\/15\/11\/2718\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T19:40:45Z","timestamp":1760125245000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2072-4292\/15\/11\/2718"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,5,24]]},"references-count":51,"journal-issue":{"issue":"11","published-online":{"date-parts":[[2023,6]]}},"alternative-id":["rs15112718"],"URL":"https:\/\/doi.org\/10.3390\/rs15112718","relation":{},"ISSN":["2072-4292"],"issn-type":[{"value":"2072-4292","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,5,24]]}}}