Abstract An attempt has been made in this paper to classify multispectral images using customized particle swam optimization. To reduce the time consumption due to increase in dimensionality of multispectral imagery a preprocessing is done using feature extraction based on decision boundary. The customized particle swam optimization then works on the reduced multispectral imagery to find globally optimal cluster centers. Here particle swam optimization is tailored for classification of multispectral images as customized particle swam optimization. The modifications are performed on the velocity function such that velocity in each iteration is updated as a factor of g-best (global best) alone and the particle structure is made to incorporate the entire cluster centers of the reduced imagery. The initialization of particles is accomplished using modified k-means in order to retain the simplicity. AVIRIS images are used as test site and it was found that the customized particle swam optimization finds the globally optimal clusters with 98.56% accuracy.

References

[1] Babu G. and Murty M., A Near-Optimal Initial Seed Selection in K -Means Algorithm Using a Genetic Algorithm, Pattern Recognition Letters , v ol. 14, pp. 763-769, 1993.

[2] Fogel D., An Introduction to Simulated Evolutionary Optimization, IEEE Transactions on Neural Networks , vol. 5, no. 1, pp. 3-14, 1994.

[3] Jones D. and Beltramo M., Solving Partitioning Problems with Genetic Algorithms, in Proceedings of the 4 th International Conference Genetic Algorithms , San Mateo, 1991.

[4] Klein R. and Dubes R., Experiments in Projection and Clustering by Simulated Annealing, Pattern Recognition , vol. 22, pp. 213-220, 1989.

[5] Lee C. and Landgrebe D., Decision Noundary Feature Selection for Nonparametric Classifiers, in Proceedings of the SPSES 44 Annual Conference , pp. 475-478, 1991.

[6] Lee C. and Landgrebe D., Feature Extraction Based on Decision Boundaries, in IEEE Transactions on Pattern Analysis and Machine Intelligence , v ol. 15, no. 4, pp. 388-400, 1993.

[7] Venkatalakshmi K. and Shalinie M., Multispectral Image Classification Using Modified k-Means Algorithm, Selected for publication in Neural Network World .

[8] Venkatalakshmi K., Sridhar S., and Shalinie M., Neuro-Statistical Classification of Multispectral Images Based on Decision Fusion, Neural Network World , vol. 16, no. 2, pp. 97-107, 2006. Venkatalakshmi Krishnan is working as a senior lecturer in the Department of Information Technology, Thiagarajar College of Engineering, Madurai, India. She has published papers in national and international conferences and international journals. Her research areas include pattern recognition, evolutionary algorithms, and multispectral data fusion. Anisha Praisy is doing final year B.Tech (IT) in Thiagarajar College of Engineering, India. She has published papers in national conference and international journals. Her areas of interest include image processing, and pattern recognition. 78 The International Arab Journal of Information Technology, Vol. 5, No. 4, October 2008 Maragathavalli R. is doing final year B.Tech (IT) in Thiagarajar College of Engineering, India. She has published papers in national conference and international journals. Her areas of interest include image processing, and evolutionary algorithms. Mercy Shalinie is working as an assistant professor in the Department of Computer Science and Engineering, Thiagarajar College of Engineering, Madurai, India. She has obtained her doctorate from Madurai Kamaraj University in the area of neuro-fuzzy systems for pattern recognition. She has to her credit, 30 papers in National, International Conferences and Journals.