Open Collections

Ye, Yinan

University of British Columbia

Polygon representations are common in environmental applications. In forestry, forest stand polygons are foundational to forest inventory and monitoring, but delineation remains a labor-intensive and subjective process when conducted manually. To address these limitations, this thesis presents an automated segmentation algorithm for delineating and updating forest stands using Landsat imagery and forest attributes. A novel two-phase segmentation algorithm was developed to integrate Landsat surface reflectance data with structural, age, and species attributes to produce segments that resemble those in conventional forest inventories. This approach was applied to five study sites totalling 45 Mha selected to represent distinct forest landscapes across Canada. Results showed that grouping micro-segments by stand height variation minimized variance within forest segments in western Canada, while stem volume minimized variance within forest segments in boreal forest sites. Stand height variation also consistently showed the lowest Global Moran's I (ranging from 0.1 to 0.2), indicating that variance between segmented forest stands was greatest and most heterogeneous from their neighbors. Interestingly, while using species produced relatively homogeneous forest stands, it was less successful in discriminating between stands. Building on this foundation, a temporally consistent segmentation method was introduced to update forest polygon boundaries over time. Using an annual Landsat time series from 1984 to 2022, the method maintained boundaries for non-disturbed / unchanged areas and dynamically re-segmented regions experiencing disturbance and regeneration. Applied to six fire-disturbed study sites in British Columbia, the algorithm captured changes in stand structure during post-disturbance recovery, with mean segment area and shape complexity gradually returning to pre-disturbance levels over 15 years. Together, these contributions provide a unified framework for generating and updating forest stand polygons in a transparent and automatic manner. The resulting polygon products support applications requiring polygon representation of changing environmental conditions and forest attributes and offers a practical alternative to proprietary segmentation tools by leveraging freely available remote sensing data and open-source methods.

Text

2025-06-26

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/91444

Forestry

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/