Predicting tropical forest stand structure parameters from Fourier transform of very high-resolution remotely sensed canopy images
Predicting stand structure parameters for tropical forests from remotely sensed data has numerous important applications, such as estimating above-ground biomass and carbon stocks and providing spatial information for forest mapping and management planning, as well as detecting potential ecological determinants of plant species distributions. As an alternative to direct measurement of physical attributes of the vegetation and individual tree crown delineation, we present a powerful holistic approach using an index of canopy texture that can be extracted from either digitized air photographs or satellite images by means of two-dimensional spectral analysis by Fourier transform. We defined an index of canopy texture from the ordination of the Fourier spectra computed for 3545 1-ha square images of an undisturbed tropical rain forest in French Guiana. This index expressed a gradient of coarseness vs. fineness resulting from the relative importance of small, medium and large spatial frequencies in the Fourier spectra. Based on 12 1-ha control plots, the canopy texture index showed highly significant correlations with tree density (R2 = 0·80), diameter of the tree of mean basal area (R2 = 0·71), distribution of trees into d.b.h. classes (R2 = 0·64) and mean canopy height (R2 = 0·57), which allowed us to produce reasonable predictive maps of stand structure parameters from digital aerial photographs. Synthesis and applications. Two-dimensional Fourier analysis is a powerful method for obtaining quantitative characterization of canopy texture, with good predictive ability on stand structure parameters. Forest departments should use routine forest inventory operations to set up and feed regional databases, featuring both tree diameter figures and digital canopy images, with the ultimate aims of calibrating robust regression relationships and deriving predictive maps of stand structure parameters over large areas of tropical forests. Such maps would be particularly useful for forest classification and to guide field assessment of tropical forest resources and biodiversity.
| Main Authors: | , , , |
|---|---|
| Format: | article biblioteca |
| Language: | eng |
| Published: |
Wiley
|
| Subjects: | U10 - Informatique, mathématiques et statistiques, K10 - Production forestière, couvert, forêt tropicale, peuplement forestier, caractéristique du peuplement, modèle mathématique, modèle de simulation, imagerie, analyse d'image, http://aims.fao.org/aos/agrovoc/c_1262, http://aims.fao.org/aos/agrovoc/c_24904, http://aims.fao.org/aos/agrovoc/c_28080, http://aims.fao.org/aos/agrovoc/c_34910, http://aims.fao.org/aos/agrovoc/c_24199, http://aims.fao.org/aos/agrovoc/c_24242, http://aims.fao.org/aos/agrovoc/c_36760, http://aims.fao.org/aos/agrovoc/c_36762, http://aims.fao.org/aos/agrovoc/c_3093, http://aims.fao.org/aos/agrovoc/c_3081, |
| Online Access: | http://agritrop.cirad.fr/530272/ http://agritrop.cirad.fr/530272/1/document_530272.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Predicting stand structure parameters for tropical forests from remotely sensed data has numerous important applications, such as estimating above-ground biomass and carbon stocks and providing spatial information for forest mapping and management planning, as well as detecting potential ecological determinants of plant species distributions. As an alternative to direct measurement of physical attributes of the vegetation and individual tree crown delineation, we present a powerful holistic approach using an index of canopy texture that can be extracted from either digitized air photographs or satellite images by means of two-dimensional spectral analysis by Fourier transform. We defined an index of canopy texture from the ordination of the Fourier spectra computed for 3545 1-ha square images of an undisturbed tropical rain forest in French Guiana. This index expressed a gradient of coarseness vs. fineness resulting from the relative importance of small, medium and large spatial frequencies in the Fourier spectra. Based on 12 1-ha control plots, the canopy texture index showed highly significant correlations with tree density (R2 = 0·80), diameter of the tree of mean basal area (R2 = 0·71), distribution of trees into d.b.h. classes (R2 = 0·64) and mean canopy height (R2 = 0·57), which allowed us to produce reasonable predictive maps of stand structure parameters from digital aerial photographs. Synthesis and applications. Two-dimensional Fourier analysis is a powerful method for obtaining quantitative characterization of canopy texture, with good predictive ability on stand structure parameters. Forest departments should use routine forest inventory operations to set up and feed regional databases, featuring both tree diameter figures and digital canopy images, with the ultimate aims of calibrating robust regression relationships and deriving predictive maps of stand structure parameters over large areas of tropical forests. Such maps would be particularly useful for forest classification and to guide field assessment of tropical forest resources and biodiversity. |
|---|