Simultaneous Estimations of Forest Parameters using Aerial Photograph Interpreted Data and the k Nearest Neighbour Method

Abstract

Information about the state of the forest is of vital importance in forest management planning. To enable high-precision modelling, many forest planning systems demand input data at the single-tree level. The conventional strategy for collecting such data is a plot-wise ® eld inventory. This is expensive and, thus, cost-ef® cient alternatives are of interest. During recent years, the focus has been on remote sensing techniques. The k nearest neighbour (kNN) estimation method is a way to assign plot-wise data to all stands in a forest area, using remotely sensed data in connection with a sparse sample of ® eld reference plots. Plot-wise aerial photograph interpretations combined with information from a stand register were used in this study. Nearness to a reference plot was decided upon using a regression transform distance. Standing stem volume was estimated with a relative root mean square error (RMSE) equal to 20% at the stand level, while age could be estimated with a RMSE equal to 15%. A cost-ef® cient data-capturing strategy could be to assign plot data with the presented kNN method to some types of forest, while using traditional ® eld inventories in other, more valuable, stands.

Localization of growth estimates using non-parametric imputation methods

“…The non-parametric approach subsumes many methods and variations on methods, among which those that are suitable for multidimensional situations include nearest neighbour methods and generalized additive models. Nearest neighbour methods are widely used in many forestry applications, including generalization of sample tree information, estimation of diameter distribution and multivariate and multisource forest inventory applications (see Korhonen and Kangas, 1997;LeMay et al, 2008;Maltamo and Kangas, 1998;Moeur and Stage, 1995;Moeur and Hershey, 1999;Holmströ m et al, 2001;Packalé n and Maltamo, 2007). Non-parametric imputation methods can be localized, for example, by reference to the spatial distance between a target tree and its neighbouring trees or with moving geographical zones.…”

Airborne Laser Scanning for the Site Type Identification of Mature Boreal Forest Stands

“…The nearest neighbor methods have been widely used for estimating continuous forest variables (e.g., by [8,33,34]) and some extent for determining discrete forest variables (e.g., [32,35]). Peuhkurinen et al [36] studied species-specific diameter distributions and saw log recoveries with the k-NN method by using first pulses of the data and noted that the method they introduced can also be applied in different classification procedures.…”

Rapid assessment of wind storm-caused forest damage using satellite images and stand-wise forest inventory data

“…For more details, see text. (Moeur & Stage 1995, Temesgen et al 2003, aerial images together with, or independently of, data derived from laser scanning (Naesset 1997, 2007, Means et al 1999, Holmström et al 2001, Muinonen et al 2001, Holmgren 2004, Eskelson et al 2008. More recently, similar research in Lithuania reached the conclusion that summary statistics only, such as the mean volume of growing stock per hectare for some areas, could be applicable operationally if the k-NN predictions were done at a forest compartment level .…”