1217
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
Figure 3. Segmentation of a Google Earth screen capture image with scale 50
4. CONCLUSIONS
In this work, a multi-resolution image segmentation method
combining spectral and shape features is proposed, with
reference to the basic idea of eCogntion. In most cases, our
segmentation method produces highly visually homogeneous
parcels in arbitrary resolution on different types of images. We
declare that our method reaches the level of eCognition in
segmenting precision, and satisfies the practical need of
segmenting remotely sensed imagery with fair algorithm
efficiency. As a first step for further analyses, multiresolution
segmentation can be used to produce image object primitives.
Starting from this, we can carry out a lot of higher level image
interpretation including image classification, information
extraction, and object recognition, etc.
REFERENCES
Aplin P., Atkinson P., Curran P., 1999. Per-field Classification
of Land Use Using the Forthcoming Very Fine Resolution
Satellite Sensors: Problems and Potential Solutions. In:
Advances in Remote Sensing and GIS Analysis. Chichester:
Wiley and Sons Inc., 219-239
Definiens website, 2007. http://www.definiens.com (access
2007-12-30)
Baatz M. and Schape A., 2007. Multiresolution Segmentation-
an Optimization Approach for High Quality Multi-scale Image
Segmentation, http://www.definiens-imaging.com (access 2007-
12-30)
Stoev S. L., 2000. RaFSi - A Fast Watershed Algorithm Based
on Rainfalling Simulation. In Proceedings of 8-th International
Conference on Computer Graphics, Visualization, and
Interactive Digital Media, Plzen City, Czech Republic
Vincent L. and Soille P., 1991. Watershed in digital spaces: an
efficient algorithm based on immersion simulations. IEEE
Trans. Patt. Anal, and Mach. Int., 13(6): 583-598.
Patrick De Smet and Rui Luis V.P.M. Pires, 2000.
Implementation and analysis of an optimized rainfalling
watershed algorithm. Proc.SPIE, 3974:759-766
ACKNOWLEDGEMENTS
This work is supported by Chinese National Programs for High
Technology Research and Development (No.2007AA12Z224,
2006AA12Z146), and startup fund of Nanjing Normal
University (No. 2006105XGQ0035)