Spectral reflectance models for spruce (Picea abies L.) damage estimation employing aerial digital data

2001

S. 220-228

3 Abb., 4 Tab., 12. Lit. Ang.

Unselbständiges Werk

200087225

The paper deals with spruce damage estimation employing spectral reflectance models. These were derived on the basis of large scale digital data. Not only original data sets were used for model construction but also some other data employing components of IHS transformation, PCA analysis and image algebra. On some color composites it is possible to distinguish spruce damage in great detail at the level of single crowns and also within them. The differences in crown morphology and color differences was considered as the principal characters in the damage detection for purposes of spectral signature collection. The crown surface spectral profiles and methods of discriminant analysis were used for selection of the best digital data sets for spruce defoliation detection. Reflectance models derived enable to estimate spruce defoliation with different accuracy (SE% = 11-50%), coefficients of determination vary in the range of 0.25-0.96. In spite of the fact that the accuracy of reflectance models is not very favorable, they are a good basis for collection of information for two-phased sampling procedures. Damage dot estimations by reflectance models are useful also for analysis and interpolation of damage with respect to spatially correlated data employing geostatistics methods. An example of both approaches is presented in this paper.