Automatic Identification of Tree Crowns in Different Topology

C.-C. Ko, C.-S. Lin, J.-P. Huang, and R.-C. Hsu (Taiwan)


Satellite sensed imagery, multi-spectral imagery, Normalized difference vegetation, Spectral angle, Morphological grayscale reconstruction


Efficient forest management requires resource information at fine scale. Satellite sensed imagery is one of potential tools for collecting forest resource. In this paper, we developed an automated delineation of tree crown based on a series of image processing procedures and spectral angle using multi-spectral information. Above all, the proposed schemes can be applied to an open forest with complicated scenes. Experimental results reveal that the proposed methods can effectively identify and isolate the tree crown compared to traditional approach, and manual delineation. It also provides detailed forest resource information, and can be used as a reliable management tool for forest resources.

Important Links:

Go Back