Automatic Identification of Individual Tree Crowns in Mixed Forests Using Fusion of LIDAR and Multispectral Data
2013
Linda Gulbe, Ints Mednieks

One of the most important steps in performing automated forest inventory at the individual tree level using remote sensing data is the identification of individual trees. The aim of this paper is to present a LIDAR and multispectral data fusion approach exploiting Template Matching (TM) method for individual tree identification and evaluate its accuracy comparing with single data source cases. Data fusion was achieved in two ways: 1) at pixel level combining data sets using principal components transform and wavelet decomposition, and 2) at feature level by combining intermediate results of the TM method. The best overall accuracy, namely 76%, was achieved by employing wavelet decomposition based data fusion; however this result does not outwork a lot the usage of LIDAR data set alone.


Atslēgas vārdi
identification of trees, multispectral imaging, LIDAR, template matching, data fusion

Gulbe, L., Mednieks, I. Automatic Identification of Individual Tree Crowns in Mixed Forests Using Fusion of LIDAR and Multispectral Data. Datorvadības tehnoloģijas. Nr.14, 2013, 93.-99.lpp. ISSN 2255-9108. e-ISSN 2255-9116.

Publikācijas valoda
English (en)
RTU Zinātniskā bibliotēka.
E-pasts: uzzinas@rtu.lv; Tālr: +371 28399196