- Tytuł:
-
Wykorzystanie lotniczej teledetekcji hiperspektralnej w klasyfikacji gatunkowej lasów strefy umiarkowanej
Airborne hyperspectral data for the classification of tree species a temperate forests - Autorzy:
-
Wietecha, M.
Modzelewska, M.
Stereńczak, K. - Powiązania:
- https://bibliotekanauki.pl/articles/987129.pdf
- Data publikacji:
- 2017
- Wydawca:
- Polskie Towarzystwo Leśne
- Tematy:
-
lesnictwo
strefa umiarkowana
teledetekcja
dane hiperspektralne
wykorzystanie
lasy
sklad gatunkowy
drzewa lesne
klasyfikacja
remote sensing
hyperspectral data
tree species classification - Opis:
- The review focuses on use of airborne hyperspectral imagery in forest species classification. Studies mentioned in the review concern hyperspectral image classification with use of various methods. Only research, where study area is located in Europe or North America were selected. Articles were reviewed with respect to used pre−processing methods, methods of feature selection or feature extraction, algorithms of image classification and trees species which were classified. The whole process of acquiring and working with hyperspectral data is described. Different approaches (e.g. use or skip atmospheric corrections) were compared. In each article, various deciduous and conifer species were classified. Studies comparing several classification algorithms (Spectral Angle Mapper, Support Vector Machine, Random Forest) were mentioned. In most cases SVM gives the best results. Species, which are classified with the highest accuracy, include Scots pine (Pinus sylvestris) and Norway spruce (Picea abies). Broadleaved species are, in general, classified with lower accuracy than conifer ones. Within broadleaved trees, European beech (Fagus sylvatica) and oaks (Quercus sp.) are classified with the highest accuracy.
- Źródło:
-
Sylwan; 2017, 161, 01; 3-17
0039-7660 - Pojawia się w:
- Sylwan
- Dostawca treści:
- Biblioteka Nauki
Artykuł