Scholarly record
APPLICATION OF FRACTAL DIMENSION AND ENTROPY AS THE MEASURES FOR RECOGNIZING OFFSHORE VEGETATION ON AERIAL PHOTOGRAPHS
Abstract
One of the certificates that characterizes ecological landscape changeability is fractal dimension. According to Environmental Protection Agency, the dynamics and vectors of those changes can be determined in three-dimensional space where fractal dimension makes one of the axes. Entropy can be regarded as another measurement of changeability (chaos). Application of such measurement of landscape changeability as comparative patterns of land cover is the subject matter of the paper. A fragment of a vertical color photograph of lake Lukajno shoreland taken by the author of the paper in September 2006 has been chosen for research. The concept of fractal dimension does not have only one definition. Hence, so numerous methods of its calculating. Proposed by Clarke method of triangular truncated prism has been applied in research. The surface of the photograph was divided into 25 by 25 pixel squares. For each of them, in channels R, G, B fractal dimension and entropy were calculated. The created mosaics represent the distribution of local fractal and entropy dimension values. Three vegetation concentrations and lake bed deprived of vegetation were chosen for land cover analysis. The chosen vegetation concentrations include osiers, elodeids (vegetation immersed in water) and heliophites (rushes). Standard values of dimension and entropy were determined for each of them as taken from 90 representative fields. In channels R, G, B mean values were calculated and variance-covariance matrix were defined. Applying Baran?s method, elements of pattern mean error ellipse were calculated for each category of cover respectively. The classification of the pixels of local value of dimension and entropy was based on the assumption that pixels whose values R, G, B are to be found inside the ellipse of a given category, belong to the category of determined probability. Two classifications were carried out, with assumed probability equal 50%, depicting them as colorful rasters. The visual estimation of the correctness of vegetation site recognition has confirmed the applicability of recognition features. In order to increase the probability of the pixel affinity do ellipse of a given class to 75%, classification was carried out in which it is assumed that a pixel belongs to an ellipse of a given class in the spaces of both recognition features. Visual analysis of the classification results has confirmed that both fractal dimension and entropy are essential recognition features of vegetal cover relating to water habitat. Erroneous classification or lack of category recognition might be caused by heterogeneity of vegetation patches and high sensitivity of the method to technical imperfection of information carrier. Vegetation mosaic configurations do not show any distinct interior limits and vegetation concentrations can not be treated as pure fractals but rather as a blend of such, namely multifractals.
Publication Impact Profile
Publication details
References9
Peitgen H. O., Jürgens H., and Saupe D., Granice chaosu fraktale. Warsaw: Wydawnictwo Naukowe PWN, 1996.
Frohn R.C., Remote sensing for landscape ecology: New metric indicators for monitoring, modelling and assessment of ecosystems. Lewis Publishers, 1998.
Mandelbrot B., ‘How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension’, Science (80-. )., vol. 156, no. 3775, pp. 636–638, 1967.
Clarke K. C., ‘Computation of the fractal dimension of topographic surfaces using the triangular prism surface method’, Comput. Geosci., vol. 12, no. 5, pp. 713–722, 1986.
De Jong S. M. and Burrough P., ‘A fractal approach to the classification of Mediterranean vegetation types in remotely sensed images’, Photogramm. Eng. Remote Sensing, vol. 61, pp. 1041–1053, 1995.
Qiu H. and Lam N., ‘Fractal characterization of hyperspectral imagery’, Photogramm. Eng. Remote Sens., vol. 65, no. 1, pp. 63–71, 1999.
Sun W., ‘Triangular Prism Metod for Computing the Fractal Dimension of Remote Sensing Image’, Photogramm. Eng. Remote Sensing, vol. 72, no. 4, pp. 373–382, 2006.
Haralick R. M., ‘Statistical and structural approaches to texture’, Proc. IEEE, vol. 67, no. 5, pp. 786–804, 1979.
Richards J. and Jia X., Remote Sensing Digital Image Analysis. 2006.
View or Download full articleAccess options
SWS access login
Login as SWS Scientific CommitteeLogin as SWS Scientific PartnerLogin as SWS AuthorAuthors and approved SWS contributors will read and export their own linked papers after identity matching by SWS profile, email and SGEM GlobalID.
For librarian assistance: [email protected]
Purchase Instant Access
- Article can be downloaded after successful payment.
- Article may be used according to SWS library access terms.
- Article cannot be redistributed.
