| Autor: |
Machala, Martin, Zejdová, Lucie |
| Předmět: |
|
| Zdroj: |
European Journal of Remote Sensing; 2014, Vol. 47 Issue 1, p117-131, 15p |
| Abstrakt: |
The objective of this research was to establish a set of rules for using Object-based Image Analysis (OBIA) to semi-automatically map a given densely forested area. Aerial images (RGB and NIR bands) and LiDAR elevation data provided the raw data and were analysed using eCognition Developer 8 software. All the non-forested areas, including built-up areas, water surfaces or agricultural land, were identified first. The forested areas were then classified as stands composed principally of broadleaf trees, coniferous trees, mixed forest or clear-cuts, which was achieved with an accuracy of almost 90%. Subsequently the stands were classified on the basis of height, using 5 metres intervals, and this was achieved with an accuracy of just over 70%. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
