land-use-land-cover-classification

Preprocessing and Classification of landsat 8 satellite images using Random Forest Classifier.

Preprocessing includes:

False color composite creation.
DN conversion to TOA Reflectance.
Atmospheric Correction using Dark Object Subtraction.
Cloud Correction using time series data.
Color Balancing using CLAHE (Contrast Local Adaptive Histogram Equalisation) and contrast stretching.
Reprojection to WGS84 coordinates.
Mosaicing and clipping using shapefile.
Rasterize training shapefiles necessary for classification.

Generation of ground truth data based on deep learning:

Unsupervised learning using autoencoders

One notebook illustrates the use of google earth engine for classification of Landsat 8 imagery using SVM classifier.

Name		Name	Last commit message	Last commit date
Latest commit History 14 Commits
README.md		README.md
Tripura.ipynb		Tripura.ipynb
classification mosaic-new.ipynb		classification mosaic-new.ipynb
decadal change.ipynb		decadal change.ipynb
grab_meta.py		grab_meta.py
image_stitching.pdf		image_stitching.pdf
lulc classification.ipynb		lulc classification.ipynb
lulc_utils.py		lulc_utils.py
manudeo river basin.ipynb		manudeo river basin.ipynb
preprocessing.ipynb		preprocessing.ipynb
preprocessing.py		preprocessing.py
training data preparation.ipynb		training data preparation.ipynb
view.ipynb		view.ipynb

Provide feedback