Canadian Forest Service Publications
Impact of sun-surface-sensor geometry upon multitemporal high spatial resolution satellite imagery. 2008. Wulder, M.A.; Ortlepp, S.M.; White, J.C.; Coops, N.C. Canadian Journal of Remote Sensing 34(5): 455-461.
Available from: Pacific Forestry Centre
Catalog ID: 29080
CFS Availability: PDF (download)
Agile high spatial resolution satellites, such as QuickBird, are capable of varying their in-track and cross-track view angles, thereby reducing the time required for the satellite to revisit the same location. However, variation in these view angles can impact the structural representation of image objects, with implications for multitemporal analyses of high spatial resolution imagery. To demonstrate this, we used four QuickBird images collected annually between 2003 and 2006 over a managed forest location near Merritt, British Columbia, Canada. Selected linear features (i.e., roads) were used to assess the geometric fidelity of the images, while forest structure, characterized with image segments and local maxima stem counts, was used to assess the impact of variability in viewing geometry. Using the 2003 QuickBird panchromatic image as a base, over 80% of the selected linear features in the 2004, 2005, and 2006 images were within less than 1 m of the 2003 control, and more than 96% were within 2m. Using pairwise t tests of the stem counts per segment (image-derived spatial units analogous to small groups of trees) for each year of imagery, we identified a significant difference between the stem count generated from the 2005 imagery and stem counts generated from each of the other images. The 2005 image had a markedly different in-track view angle and satellite azimuth relative to the other images considered, and therefore, although our time series of QuickBird imagery had demonstrated geometric fidelity, the forest structural information extracted from our time series was not consistent. We conclude that careful attention should be paid to these image acquisition parameters when selecting high spatial resolution images for multitemporal analyses in forested environments.
- Date modified: