Open Collections

Using texture energy measures for the segmentation of forest scenes

University of British Columbia

The stratification of forest cover is a basis for most forest management activities. With the development of computer-based image analysis systems, attempts have been made to automate the photo-interpretation process using local spectral signatures. The results have generally failed to meet the classification accuracy of trained humans. To complement the spectral analysis, local texture analysis is proposed. This thesis investigates the potential of a technique based on texture energy measures (Laws 1980) for the segmentation of forest scenes. The technique is tested with a set of two panchromatic aerial photographs digitized at four spatial resolutions: 1.5, 5, 10 and 20 m. The main conclusion of these tests is that given an adequate spatial resolution, texture energy measures can provide a reliable segmentation of forest cover into two classes of textures corresponding to stands with open and closed canopies. Only the finest spatial resolution (1.5 m) proved to be adequate.

Text

2010-06-14

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/25736

Forestry

University of British Columbia

Graduate