Individual tree crown (ITC) techniques


Supervised classification of automatically delineated tree crowns at different spatial resolutions

A 70 cm/pixel multispectral image

Green = Red pine
Yollow = White pine
Larger image [16 kb GIF]

A 100 cm/pixel multispectral image

Blue = Red spruce
Red = White spruce
Larger image [24 kb GIF]

 

Multispectral supervised classification into four coniferous species of tree crowns automatically delineated on:

  • a 70 cm/pixel multispectral image (a)
  • a 100 cm/pixel multispectral image (b)

The image used in (b) was generated from the first image (a), simulating "very" approximatively an image from the next generation of earth observation satellites (e.g., IKONOS).

Average classification accuracies are typically around 74-81% for four coniferous species, depending on the spatial resolution (36 to 100 cm/pixel) (Gougeon, 1995a).

Work with digitized colour infrared aerial photography

(a)Colour infrared aerial photograph (1:90 000) of the China Creek region on Vancouver Island, British Columbia, Canada, digitized to a spatial resolution of ~60 cm/pixel. Click here to see a larger image [86 kb GIF]
(a)Colour infrared aerial photograph (1:90 000) of the China Creek region on Vancouver Island, British Columbia, Canada, digitized to a spatial resolution of ~60 cm/pixel.
Larger image [86 kb GIF]
(b)Isolation, delineation and classification of the tree crowns detected in image (a) (Gougeon, 1996). Click here to see a larger image [25 kb GIF]
(b)Isolation, delineation and classification of the tree crowns detected in image (a) (Gougeon, 1996).
Larger image [25 kb GIF]

 

 


Project status

  • On-going