#K-mean cluster of Gandhinagar SRTM 30m DEM Using SCIPY Lib

import scipy as sp
#reading rasterdata into array form "osgeo"
from osgeo import gdal
#plot graph
from pylab import plot,show
import numpy as np
from scipy.cluster.vq import kmeans,vq

dem=gdal.Open(r"C:\Users\lenovo\Desktop\scipy\SRTM-30.tif")
features = np.array(dem.GetRasterBand(1).ReadAsArray(),dtype=float)

print (features.shape)
print (features)

# computing K-Means with K = 2 (2 clusters)
codes = 2

# assign each sample to a cluster
centroids,_ = kmeans(features,codes)
#centroids will show me the important pattern of data

#vq= vector quantization
#vq vector quantaization =Assign codes from a code book to observations.
idx,_ = vq(features,centroids)

# some plotting using numpy's logical indexing
plot(features[idx==0,0],features[idx==0,1],'ob',
features[idx==1,0],features[idx==1,1],'or')
plot(centroids[:,0],centroids[:,1],'sg',markersize=8)
show()