MACHINE LEARNING APPROACHES FOR BURNED AREA IDENTIFICATION USING SENTINEL-2 IN CENTRAL KALIMANTAN
Abstract
Šumski ili kopneni požar je katastrofa koja ima veliki utecaj na životnu sredinu. Indonezija se svake godine suočava sa šumskim i kopnenim požarima, većinom u Kilimantanu i Sumateri. Optički daljinski satelit sa senzorima postaje obećavajuća tehnologija koja se u okviru sistema kontrole prirodnih katastrofa može koristiti za brzu identifikaciju spaljenih područja. Korišćenje mašinskog učenja za brzu identifikaciju spaljenog područja je u porastu s obzirom da se može koristiti za automatsku identifikaciju spaljenih područja na velikom prostoru. Ova studija je u junu i avgustu 2019 godine izvršila procenu korišćenja vektorske mašine za podršku (SVM), nasumične šume (RF) i duboke neuralne mreže (DNN) u pokrajini Centralni Kalimantan kao slučajeve pre i posle požara, koristeći slike Sentinel-2. Na tim klasifikatorima su korišćene neuravnotežene i uravnotežene grupe podataka s različitim hiper-parametrima. Korišćeni su i podaci o hotspot dobiveni iz MODIS-a i Suomi NPP podataka kao grupe podataka za obučavanje i testiranje. Na osnovu ove studije, grupa neuravnoteženih podataka utiče na vrednosti vezane za preciznosti i opoziv kao i na tačnost SVM i DNN klasifikatora, tako da, u slučaju vrednosti vezanih za preciznost, opoziv iIi tačnost, RF nadmašuje SVM i DNN metode. Ovo se pokazuje kroz RF metod koji ne prolazi kroz značajne promene u tim vrednostima i uravnotežene i neuravnotežene grupe podataka. Međutim, visoka tačnost se još uvek može postići putem SVM, RF i DNN metoda s neuravnoteženim ili uravnoteženim grupama podataka.
Forest or land fire is a disaster that has a large impact on the environment. Every year, Indonesia undergoes forest or land fire mainly in Kalimantan and Sumatera. Optical remote sensing satellite becomes a promising technology that can be utilized to identify the burned area in quick time for disaster management response. The use of machine learning for burned area identification is rising since it can be used to identify the burned area in a vast area automatically. This study evaluated the use of Support Vector Machine (SVM), Random Forest (RF), and Deep Neural Network (DNN) in the Central Kalimantan province on June and August 2019 as pre-fire event and post-fire event using Sentinel-2 imageries. An imbalanced and a balanced dataset with varying hyper-parameter were used on those classifiers. Hotspot data derived from MODIS and Suomi NPP data are also used as training and testing dataset. Based on the study, the imbalanced dataset influences precision and recall values, as well as the accuracy of SVM and DNN classifier, so that RF outperforms SVM and DNN methods in case of precision and recall values, as well as accuracy. This is shown through RF method that is relatively not experiencing significant changes on those values in both an imbalanced or balanced dataset. However, the high accuracy is still can be achieved by SVM, RF, and DNN methods with an imbalanced or a balanced dataset.
References
2. Suwarsono., Rokhmatuloh., Waryono, T. (2013). Pengembangan model identifikasi daerah bekas kebakaran hutan dan lahan (burned area) menggunakan citra MODIS di Kalimantan. Jurnal Penginderaan Jauh, vol.10, 93–112
3. Zubaidah, A., Vetrita, Y., Khomarudin, M. R. (2014). Validasi hotspot MODIS di wilayah sumatera dan kalimantan berdasarkan data penginderaan jauh SPOT-4 tahun 2012. Jurnal Penginderaan Jauh vol. 11, 1–15
4. The World Bank. (2016). Laporan pengetahuan lanskap berkelanjutan Indonesia: Kerugian dari kebakaran hutan. Jakarta
5. Ministry of Environment and Forestry Republic of Indonesia. Rekapitulasi luas kebakaran hutan dan lahan (Ha) per provinsi di Indonesia tahun 2014-2019, from: http://sipongi.menlhk.go.id/hotspot/luas_kebakaran, accessed 2019-11-01
6. Pinem, T. (2016). Kebakaran hutan dan lahan gambut: kajian teologi ekofeminisme. Gema Teologi, vol.1, no. 2, 139-166
7. Yulianti, N. (2018) Pengenalan bencana kebakaran dan kabut asap lintas batas, 1st ed., IPB Press: Bogor
8. Chuvieco, E., Mouillot, F., van der Werf, G. R., San Miguel, J., Tanasse, M., Koutsias, N., García, M., Yebra, M., Padilla, M., Gitas, I., Heil, A., Hawbaker, T. J., Giglio, L. (2019). Historical background and current developments for mapping burned area from satellite Earth observation. Remote Sensing of Environment, vol. 225, 45–64, DOI: 10.1016/j.rse.2019.02.013
9. Fanin, T., Van Der Werf, G. R. (2015). Relationships between burned area, forest cover loss, and land cover change in the Brazilian Amazon based on satellite data. Biogeosciences, vol. 12, no.20, 6033–6043, DOI: 10.5194/bg-12-6033-2015
10. Chuvieco, E., Lizundia-Loiola, J., Pettinari, M.L., Ramo, R., Padilla, M., Tansey, K., Mouillot, F., Laurent, P., Storm, T., Heil, A., Plummer, S. (2018). Generation and analysis of a new global burned area product based on MODIS 250m reflectance bands and thermal anomalies. Earth System Science Data, vol.10, 2015–2031, DOI: 10.5194/essd-2018-46
11. Langford, Z. L., Kumar, J., Hoffman, M.F. (2018). Wildfire mapping in interior alaska using deep neural networks on imbalanced datasets. 2018 IEEE International Conference on Data Mining Workshops (ICDMW)
12. Mallinis, G., Koutsias, N. (2012). Comparing ten classification methods for burned area mapping in a Mediterranean environment using Landsat TM satellite data. International Journal of Remote Sensing, vol.33, no. 14, 4408–4433, DOI: 10.1080/01431161.2011.648284
13. Bastarrika, A., Alvarado, M., Artano, K., Martinez, M.P., Mesanza, A., Torre, L., Ramo, R., Chuvieco, E. (2014) BAMS: a tool for supervised burned area mapping using landsat data. Remote Sensing, vol. 6, no. 12, 12360–12380, DOI: 10.3390/rs61212360
14. Roy, D.P., Huang, H., Boschetti, L., Giglio, L., Yan, L., Zhang, H.H., Li, Z. (2019). Landsat-8 and Sentinel-2 burned area mapping - a combined sensor multi-temporal change detection approach. Remote Sensing of Environment, vol. 231, DOI: 10.1016/j.rse.2019.111254
15. Verhegghen, A., Eva, H., Ceccherini, G., Achard, F., Gond, V., Gourlet-Fleury, S., Cerutti, P. O. (2016) The potential of sentinel satellites for burnt area mapping and monitoring in the Congo Basin forests. Remote Sensing, vol.8, no.12, DOI: 10.3390/rs8120986
16. Amos, C., Petropoulos, G. P., Ferentinos, K. P. (2019). Determining the use of Sentinel-2A MSI for wildfire burning & severity detection. International Journal of Remote Sensing, vol. 40, no. 7, 905–930, no. 3, DOI: 10.1080/01431161.2018.1519284
17. Filipponi, F. (2018). BAIS2: burned area index for Sentinel-2. Proceedings, vol. 2, DOI: 10.3390/ecrs-2-05177
18. Filipponi, F. (2019). Exploitation of Sentinel-2 time series to map burned areas at the national level: a case study on the 2017 Italy wildfires. Remote Sensing, vol.11, no. 6, DOI: 10.3390/rs11060622
19. Roteta, E., Bastarrika, A., Padilla, M., Storm, T., Chuvieco, E. (2019). Development of a Sentinel-2 burned area algorithm: Generation of a small fire database for sub-Saharan Africa. Remote Sensing of Environment, vol. 222, 1-17, DOI: 10.1016/j.rse.2018.12.011
20. Jozdani, S. E., Johnson, B. A., Chen, D. (2019). Comparing deep neural networks, ensemble classifiers, and support vector machine algorithms for object-based urban land use/land cover classification. Remote Sensing, vol. 11, no. 14, DOI: 10.3390/rs11141713
21. Noi, P. T., Kappas, M. (2018). Comparison of random forest, k-nearest neighbor, and support vector machine classifiers for land cover classification using sentinel-2 imagery. Sensors, vol. 18, no. 1, 10.3390/s18010018.
22. Maxwell, A. E., Warner, T. A., Fang, F. (2018). Implementation of machine-learning classification in remote sensing: an applied review. International Journal of Remote Sensing, vol. 39, no. 9, 2784–2817, DOI: 10.1080/01431161.2018.1433343
23. Zhu, X. X., Tuia, D., Mou, L., Xia, G. S., Zhang, L., Xu, F., Fraundorfer, F. (2017). IEEE Geoscience and Remote Sensing Magazine. October 2017, p. 1–60, DOI: 10.1109/MGRS.2017.2762307
24. Zhang, X., Chen, G., Wang, W., Wang, Q., Dai, F. (2017). Object-based land-cover supervised classification for very-high-resolution UAV images using stacked denoising autoencoders. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 10, no. 7, 3373–3385, DOI: 10.1109/JSTARS.2017.2672736
25. Pereira, A. A., Pereira, J. M. C., Libonati, R., Oom, D., Setzer, A.W., Morelli, F., Machado-Silva, F., de Carvalho, L. M. T. (2017). Burned area mapping in the Brazilian savanna using a one-class support vector machine trained by active fires. Remote Sensing, vol. 9, no. 11, DOI: 10.3390/rs9111161
26. Ramo, R., Chuvieco, E. (2017). Developing a random forest algorithm for MODIS global burned area classification. Remote Sensing, vol. 9, no. 11, DOI: 10.3390/rs9111193
27. Ramo, R., García, M., Rodríguez, D., Chuvieco, E. (2018). A data mining approach for global burned area mapping. International Journal of Applied Earth Observation and Geoinformation, vol. 73, 39–51, DOI: 10.1016/j.jag.2018.05.027
28. de Carvalho, N. S., Ferreira, I. J. M., Körting, T. S., Eduardo, L., Aragão, C.D., Anderson, L. O. (2018). Random forest and support vector machine applied for mapping burned areas in Amazon. Proceedings of XIX Brazilian Symposium on Remote Sensing p. 2833–2836, ISBN: 978-85-17-00097-3
29. Mallinis, G., Mitsopoulos, I., Chrysafi, I. (2018). Evaluating and comparing Sentinel 2A and Landsat-8 operational land imager (OLI) spectral indices for estimating fire severity in a mediterranean pine ecosystem of Greece. GIScience & Remote Sensing, vol. 55, no. 1, 1–18, DOI: 10.1080/15481603.2017.1354803
30. Pepe, M., Parente, C. (2018). Burned area recognition by change detection analysis using images derived from Sentinel-2 satellite: the case study of Sorrento Peninsula,Italy. Journal of Applied Engineering Science, vol. 16, no. 2, 225–232, DOI: 10.5937/jaes16-17249
31. ESA. (2015). Sentinel-2 user handbook, European Space Agency
32. Lasaponara, R., Tucci, B., Ghermandi, L. (2018). On the use of satellite Sentinel 2 data for automatic mapping of burnt areas and burn severity. Sustainability, vol. 10, no. 11, DOI: 10.3390/su10113889
33. Huang, C., Davis, L.S., Townshend, J.R. G. (2002). An assessment of support vector machines for land cover classification. International Journal of Remote Sensing, vol. 23, no. 4, 725–749, DOI: 10.1080/01431160110040323
