Post-hurricane forest mapping in Bory Tucholskie (Northern Poland) using random forest-based up-scaling approach of als and photogrammetry- based chm to kompsat-3 and planetscope imagery
Piotr Wężyk , Paweł Hawryło , Karolina Zięba-Kulawik
AbstractAssessing the extent of hurricane damage in forest areas is a difficult task in case of field-based inventory. In this context, remote sensing technologies are an attractive alternative to fast, inexpensive, and objective mapping of forest damage. The huge hurricane took place in Bory Tucholskie (Poland) on the night of 11/12 August 2017, at the belt with a length of approx. 300 km. The main goal of the study was to determine the suitability of PlanetScope (Dove) and KOMPSAT-3 satellite imageries for post-hurricane inventory of forest damage. The differences in Canopy Height Models (CHM; 1.0 m GSD) generated from the pre-hurricane ALS-based point clouds (density 4 pts/m2) and post-hurricane aerial photos-derived point clouds (RGB; 0.15 m GSD) have been used as reference data for forest damage degree assessment. That has been determined using continuous scale ranging from 0.0 (no damages) to 100.0 (complete damage; 100%). Predictive models of forest damage degree were built at image segment and forest stand levels using the Random Forests method. The mean values of KOMPSAT-3 as well as PlanetScope spectral bands (NIR, Red, Green and Blue) and NDVI were used as predictor variables. RMSE for predicted damage degree values calculated at stand level based on KOMPSAT-3 and PlanetScope imagery amounted to 8.0% (R2 = 0.81) and 7.1% (R2 = 0.82) accordingly. The obtained results indicate that posthurricane forest damage can be reliably and efficiently up-scaled from limited local area where precise reference data (like aerial photos) is available to wider areas using high resolution satellite images and Random Forest regression.
|Publication size in sheets||0.5|
|Book||Chirici Gherardo , Gianinetto Marco (eds.): Earth observation advancements in a changing world, Trends in earth observation, vol. 1, 2019, Associazione Italiana di Telerilevamento (AIT), ISBN 978-88-944687-1-7, 218 p., DOI:10.978.88944687/17|
|Keywords in English||Airborne laser scanning, Forest damage, Nanosatellites, LiDAR point cloud, Remote sensing|
|License||Journal (articles only); author's original; ; after publication|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.