def classify(outputClumps):
# define output class column
# rebalance the training data
classesIntCol = 'ClassInt'
rsgislib.classification.classratutils.balanceSampleTrainingRandom(
outputClumps, classesIntCol, 'classesIntColBal', 50, 5000)
classesIntCol = 'classesIntColBal'
# run the classifier
classratutils.classifyWithinRAT(outputClumps,
classesIntCol,
classesNameCol,
variables,
classifier=classifier,
classColours=None,
preProcessor=MaxAbsScaler(),
outColInt=outColInt)
# define the output colours
classColours = dict()
classColours['Other'] = [212, 125, 83]
classColours['Water'] = [157, 212, 255]
classColours['VegWater'] = [191, 255, 0]
# define variables for the classification
variables = [
'VVMin', 'VHMin', 'VVdivVHMin', 'VVMax', 'VHMax', 'VVdivVHMax',
'VVAvg', 'VHAvg', 'VVdivVHAvg', 'VVStd', 'VHStd', 'VVdivVHStd'
]
# run the classification
classratutils.classifyWithinRAT(outputClumps,
classesIntCol,
classesNameCol,
variables,
classifier=classifier,
classColours=classColours,
preProcessor=MaxAbsScaler())
############################################################################################ # Random forests ranks parameters in terms of importance to the classifier
# there is a lot which can be learnt here...
print('Feature importance: ' + inputImg)
featImportances = classifier.feature_importances_
featIndices = numpy.argsort(featImportances)[::-1]
# Print the feature ranking
print("Feature ranking:")
with open('./Importance/importance_' + inputImg + '.txt', 'w') as txt:
for f in range(len(variables)):
print("\t{0}. {1} ({2})".format(f + 1, variables[featIndices[f]],
featImportances[featIndices[f]]))
classColours['Water'] = [157, 212, 255]
classColours['Agriculture'] = [255, 255, 166]
classColours['Grassland'] = [200, 255, 187]
classColours['Forest'] = [132, 200, 76]
classColours['Urban'] = [200, 200, 200]
# define input variables
variables = [
'coastAvg', 'blueAvg', 'greenAvg', 'redAvg', 'NIRAvg', 'SWIR1Avg',
'SWIR2Avg', 'VVStd', 'wbiAvg', 'ndviAvg'
]
# run the classification
classratutils.classifyWithinRAT(clumps,
classesIntCol,
classesNameCol,
variables,
classifier=classifier,
classColours=classColours)
# export rat column to image
outimage = '/Users/Andy/Documents/Tanzania/WB_Mapping/land_cover/Landsat8_composite_clumps_classified.tif'
gdalformat = 'GTiff'
datatype = rsgislib.TYPE_8INT
fields = ['OutClass']
rastergis.exportCols2GDALImage(clumps, outimage, gdalformat, datatype, fields)
# export as shapefile
inputImg = outimage
outShp = '/Users/Andy/Documents/Tanzania/WB_Mapping/land_cover/Landsat8_composite_clumps_classified_v2.shp'
# rsgislib.vectorutils.polygoniseRaster(inputImg, outShp, imgBandNo=1, maskImg=None, imgMaskBandNo=1)
