Estimation of Mediterranean forest attributes by the application of k‐NN procedures to multitemporal Landsat ETM+ images

Geostatistical Estimation of Biomass Stock in Chilean Native Forests and Plantations

“…The main approach to overcome this problem is to include further co-variables independent to reflectance values. The k-NN method has become popular for forest inventory mapping and applications of estimation over the last years Bafetta et al, 2009;Tomppo et al, 2009;Maselli et al, 2005;Tomppo and Halme, 2004;Thessler et al, 2008). The basic idea of this method is to estimate a target attribute of an object, i.e.…”

K-NN FOREST: a software for the non-parametric prediction and mapping of environmental variables by thek-Nearest Neighbors algorithm

“…As can be easily understood, neither ED or MD emphasizes the relationship of the feature space variables to the response variable. In order to perform this task, modified forms of multidimensional distances have been proposed aiming at giving preferential consideration to the most informative feature space variables [Holmström et al, 2001;Maselli et al, 2005]. The distance weighted with fuzzy weights or Fuzzy Distance (FD) is a modification of MD where the variance-covariance matrix is computed via a fuzzy approach.…”

Experimenting the design-based k-NN approach for mapping and estimation under forest management planning

“…The exploited information from remotely sensed data are usually the DNs of the spectral bands (and/or their combination to produce vegetation indices, e.g., Maselli et al 2005) which are available for all the N pixels in the area, while the values of the Y-variable of interest (the forest attribute) are known only for the sample of n pixels corresponding to the field inventory units (each assumed to represent one pixel), characterized as the reference set. The mapping procedure is based on the nonparametric prediction of the values of Y for the pixels that do not correspond to the field inventory units, characterized as the target set.…”

“…The definition of the best configuration (i.e., the optimal setting of k-NN parameters) for the spazialization of the considered forest attributes was performed by a leave-one-out heuristic procedure where the prediction accuracy is tested against several k values (from 1 to 15) and three different distances (Euclidean, Mahalanobis, fuzzy): for detailed explanation of the procedure and the meaning of such distances, see Maselli et al (2005). The best k-NN algorithm configuration proved to be that with the TM2 band DN, the SR and RVI indices as auxiliary variables to spatialize AGB and that with the six bands of the TM image to spatialize CAI.…”