Wooded hedgerows characterization in rural landscape using very high spatial resolution satellite images

The objective of this study is to evaluate the very high spatial resolution satellite images capacity to detect and characterize the hedgerows network. The significant qualitative attributes to characterize hedgerows are composition, morphology and spatial arrangement of small elements. Remote sensing images Spot 5 and Kompsat, respectively 5m and 1m spatial resolution, were used. We applied an object-based image analysis method. The first step consists on a multi-scale segmentation, and the second step consists on a multi-criterion classification. Then, the characterization consists on a combination of shape values on the fields boundaries. Results shows that both Spot 5 and Kompsat image allows to detect automatically with a good precision the hedgerow network (84.5% and 97% respectively). Only the Kompsat image is enable to detect the finest elements of hedgerows. The very high spatial resolution image, less than 1m, allows to characterize the hedgerow cover quality, the continuity and discontinuity. This study highlights an efficient, reliable and generic method. Moreover, this characterization of landscape structures elements allows to affine the knowledge of ecological elements.