Staff Members - Benha University
GENETIC ALGORITHMS OPTIMIZATION FOR LANDCOVER CLASSIFICATION FROM HIGH RESOLUTION DIGITAL IMGERY
CERM • 2014
Publication Information
Authors
Mahmoud Al-Nokrashy Osman, Adel Ahmed Esmat, Mahmoud Salah Mahmoud, Ahmed Mohamed Hamdy
Keywords
Digital Imagery, Unsupervised Classification, Genetic Algorithm, K-means
Index.
Journal
CERM
Publisher
Not Available
Volume
36
Issue
4
Pages
319-331
publication.type
Local
Paper Link
Not Available
Supplementary Materials
Not Available
Abstract
Study of urban environmental areas involved with the use of digital imagery data has raised great
interest among researchers. High resolution imagery present difficulties for automatic
classification process due to the high spectral and spatial heterogeneity for the same class. Thus,
new concepts and techniques have been used for mapping urban areas. In this study Genetic
Algorithms (GAs) were applied to determine the optimal input parameters based on k-means
classifier as a fitness function. To assess the efficacy of the methodology and ensure the accuracy
of the product the steps undertaken in this study were subject to quality control. The best results
were obtained in the case of Population size 100 with mutation probability 0.05 with overall
accuracy of 68.89%.
interest among researchers. High resolution imagery present difficulties for automatic
classification process due to the high spectral and spatial heterogeneity for the same class. Thus,
new concepts and techniques have been used for mapping urban areas. In this study Genetic
Algorithms (GAs) were applied to determine the optimal input parameters based on k-means
classifier as a fitness function. To assess the efficacy of the methodology and ensure the accuracy
of the product the steps undertaken in this study were subject to quality control. The best results
were obtained in the case of Population size 100 with mutation probability 0.05 with overall
accuracy of 68.89%.