Identification and delineation of individual tree crowns using Lidar and multispectral data fusion

Abstract

The aim of this study is to propose and evaluate methodology for identification and delineation of individual tree crowns, using Lidar and multispectral data fusion. Methods implementing data fusion are based on image binarization (thresholding) and region growing segmentation algorithm. Results were compared with template matching method for multi-spectral data and region growing algorithm using just one data source. For a sample set, data fusion approach provided 78 % tree identification accuracy with 5 false positives, but template matching provided 72 % accuracy with 25 false positives. Data fusion approach in the delineation helped to overcome illumination effects.