Satellite Image Classification1

Category: Education

Presentation Description

No description available.


By: sakissakis (43 month(s) ago)


Presentation Transcript

Unsupervised Classification on Satellite image : 

Unsupervised Classification on Satellite image EXTERNAL GUIDE Dr S.N OMKAR Dept Of Aero Space Engineering Indian Institute of Science Bangalore INTERNAL GUIDE Prof K.R Shyalaja Dept of Computer Science Dr Ambedkar Institute of Technology Bangalore

Slide 2: 

LISS-4 Modis image of kolar dist RAW SATELLITE IMAGE

Aim of the Project Work : 

Aim of the Project Work Satellite Image Classification Input: Satellite image of low Resolution of 1501x1501 Aim: Classification of Land cover and water bodies using unsupervised clustering based algorithms Additional Aim: Reclassification of classified area Output: labeled classified image Performance analysis of classification

Slide 4: 

Input Unclassified Satellite Image K-Means Clustering Unsupervised Classification Mean Shift Clustering ? Output Classified satellite Image Re- Classification Of Classified region Comparison & Performance analysis Of different Technique

Loop Holes of Existing System : 

Loop Holes of Existing System Only primitive methods are used in unsupervised classification of land cover & water bodies Most of the work using mean shift is on Medical and simpler images. Various work has not yet proved in a completely satisfactory way the competitiveness of satellite based methods compared with ground measures and aerial surveys

Proposed System : 

The Proposed Methodology has been broadly divided Satellite Image Classification using Clustered based advanced unsupervised clustering method/algorithms like K-means , Mean shift..etc Mean shift clustering has not been used on satellite image for classification till now, according to literature survey. Reclassification of classified subsets Comparison & Performance/Accuracy evaluation of different techniques Based on evaluation results planning to go for Optimization of classification algorithms using biological inspired technique like PSO (Particle swarm optimization) after first stage of project implementation Proposed System

Slide 7: 

Original Satellite Image Classified Satellite Image Sample Input and output

Reference : 

1. Richard O. Duda, Peter E. Hart, David G. Stork, Pattern Classification, Wiley, 2002 2. K. Fukunaga, Introduction to Statistical Pattern Recognition, 2nd edition, Academic Press, London. 1990 3. K. Fukunaga, L.D. Hostetler, The estimation of the gradient of a density function with applications in pattern recognition, IEEE Trans. Inf. Theory 21 (1975) 32–40. 4. Y. Cheng, Mean shift, mode seeking, and clustering, IEEE Trans. Pattern Anal. Mach. Intell. 17 (1995) 790–799. 5. D. Comaniciu, P. Meer, Mean shift: a robust approach toward feature space analysis, IEEE Trans. Pattern Anal. Mach. Intell. 24 (2002) 603–619. 6. K.L. Wu, M.S. Yang, Alternative c-means clustering algorithms, Pattern Recognition 35 (2002) 2267–2278. 7. C.V. Stewart, Minpran: a new robust estimator for computer vision, IEEE Trans. Pattern Anal. Mach. Intell. 17 (1995) 925–938. 8. Kuo-Lung Wu, Miin-shen Yang, Mean shift-based clustering, Science direct 2007 9. Uttam kumar and T.V. Ramachandra, Endmember discrimination in MODIS using spectral angle mapper and maximum likelihood algorithms, Spectrum science Inc 2008 10. Uttam Kumar, Ramachandra T.V, Norman Kerle, Clement Atzberger, Milap Punia Evaluation of Algorithms for land cover analysis using hyperspectral data, Technical report, 2008 Reference