Evaluation of Sentinel-2 Composited Mosaics and Random Forest Method for Tree Species Distribution Mapping in Suburban Areas of Kyiv City, Ukraine

M. Kutia, V. Myroniuk, A. J. Sarkissian

2018

Abstract

The availability of comprehensive and cost-effective information on the state of suburban forests and their protection and rational use is necessary for policymakers and urban planners to make informed decisions. The lack of this information can be problematic in their efforts to develop sustainable infrastructure and the local economy while improving the environment and general wellbeing of inhabitants. The recently launched Sentinel-2 satellite presents us with the potential for improving inventory and monitoring of suburban forests. We used spatial and spectral powers of this remote sensing data through a Google Earth Engine platform for creating cloudless seasonal mosaics of Sentinel-2 imagery for the year 2015. The non-parametric Random Forest classification algorithm was used for creating continuous dominant tree species composition raster map of Kyiv suburban forests. The developed methodology, including data collection and analysis showed substantial time-savings compared to traditional inventory methods while achieving high accuracy of trees species mapping (97.8 % of overall accuracy). In summary, our approach could be considered as an application that would greatly satisfy inventory, monitoring and sustainable management of suburban forests.

Paper Citation

in Harvard Style

Kutia M., Myroniuk V. and Sarkissian A. (2018). Evaluation of Sentinel-2 Composited Mosaics and Random Forest Method for Tree Species Distribution Mapping in Suburban Areas of Kyiv City, Ukraine.In Proceedings of the International Workshop on Environmental Management, Science and Engineering - Volume 1: IWEMSE, ISBN 978-989-758-344-5, pages 597-604. DOI: 10.5220/0007563505970604

in Bibtex Style

@conference{iwemse18,
author={M. Kutia and V. Myroniuk and A. J. Sarkissian},
title={Evaluation of Sentinel-2 Composited Mosaics and Random Forest Method for Tree Species Distribution Mapping in Suburban Areas of Kyiv City, Ukraine},
booktitle={Proceedings of the International Workshop on Environmental Management, Science and Engineering - Volume 1: IWEMSE,},
year={2018},
pages={597-604},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007563505970604},
isbn={978-989-758-344-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the International Workshop on Environmental Management, Science and Engineering - Volume 1: IWEMSE,
TI - Evaluation of Sentinel-2 Composited Mosaics and Random Forest Method for Tree Species Distribution Mapping in Suburban Areas of Kyiv City, Ukraine
SN - 978-989-758-344-5
AU - Kutia M.
AU - Myroniuk V.
AU - Sarkissian A.
PY - 2018
SP - 597
EP - 604
DO - 10.5220/0007563505970604