geetools_VH/read_images.py

#==========================================================#
#==========================================================#
# Extract shorelines from Landsat images
#==========================================================#
#==========================================================#


#==========================================================#
# Initial settings
#==========================================================#

import os
import numpy as np
import matplotlib.pyplot as plt
import ee
import pdb

# other modules
from osgeo import gdal, ogr, osr
import pickle
import matplotlib.cm as cm
from pylab import ginput
from shapely.geometry import LineString

# image processing modules
import skimage.filters as filters 
import skimage.exposure as exposure
import skimage.transform as transform
import sklearn.decomposition as decomposition
import skimage.measure as measure
import skimage.morphology as morphology

# machine learning modules
from sklearn.model_selection import train_test_split
from sklearn.neural_network import MLPClassifier
from sklearn.preprocessing import StandardScaler, Normalizer 
from sklearn.externals import joblib

# import own modules
import functions.utils as utils
import functions.sds as sds

# some other settings
np.seterr(all='ignore') # raise/ignore divisions by 0 and nans
plt.rcParams['axes.grid'] = True
plt.rcParams['figure.max_open_warning'] = 100
ee.Initialize()

#==========================================================#
# Parameters
#==========================================================#

sitename = 'NARRA'

cloud_thresh = 0.7      # threshold for cloud cover
plot_bool = False       # if you want the plots
output_epsg = 28356     # GDA94 / MGA Zone 56
buffer_size = 7         # radius (in pixels) of disk for buffer (pixel classification)
min_beach_size = 20     # number of pixels in a beach (pixel classification)
dist_ref = 100          # maximum distance from reference point
min_length_wl = 200     # minimum length of shoreline LineString to be kept 

output = dict([])

#==========================================================#
# Metadata
#==========================================================#

filepath = os.path.join(os.getcwd(), 'data', sitename)
with open(os.path.join(filepath, sitename + '_metadata' + '.pkl'), 'rb') as f:
    metadata = pickle.load(f)
    
#%%
#==========================================================#
# Read S2 images
#==========================================================#

satname = 'S2'
dates = metadata[satname]['dates']
input_epsg = metadata[satname]['epsg']

# path to images
filepath10 = os.path.join(os.getcwd(), 'data', sitename, satname, '10m')
filenames10 = os.listdir(filepath10)
filepath20 = os.path.join(os.getcwd(), 'data', sitename, satname, '20m')
filenames20 = os.listdir(filepath20)
filepath60 = os.path.join(os.getcwd(), 'data', sitename, satname, '60m')
filenames60 = os.listdir(filepath60)
if (not len(filenames10) == len(filenames20)) or (not len(filenames20) == len(filenames60)):
    raise 'error: not the same amount of files for 10, 20 and 60 m'
N = len(filenames10)

# initialise variables
cloud_cover_ts = []
acc_georef_ts = []
date_acquired_ts = []
filename_ts = []
satname_ts = []
timestamp = []
shorelines = []
idx_skipped = []

spacing = '=========================================================='
msg = ' %s\n %s\n %s' % (spacing, satname, spacing)
print(msg)

for i in range(N):
    
    # read 10m bands
    fn = os.path.join(filepath10, filenames10[i])
    data = gdal.Open(fn, gdal.GA_ReadOnly)
    georef = np.array(data.GetGeoTransform())
    bands = [data.GetRasterBand(k + 1).ReadAsArray() for k in range(data.RasterCount)]
    im10 = np.stack(bands, 2)
    im10 = im10/10000 # TOA scaled to 10000
    
    # if image is only zeros, skip it
    if sum(sum(sum(im10))) < 1:
        print('skip ' + str(i) + ' - no data')
        idx_skipped.append(i)
        continue        
    
    nrows = im10.shape[0]
    ncols = im10.shape[1]
    
    # read 20m band (SWIR1)
    fn = os.path.join(filepath20, filenames20[i])
    data = gdal.Open(fn, gdal.GA_ReadOnly)
    bands = [data.GetRasterBand(k + 1).ReadAsArray() for k in range(data.RasterCount)]
    im20 = np.stack(bands, 2)
    im20 = im20[:,:,0]
    im20 = im20/10000 # TOA scaled to 10000
    im_swir = transform.resize(im20, (nrows, ncols), order=1, preserve_range=True, mode='constant')
    im_swir = np.expand_dims(im_swir, axis=2)
    
    # append down-sampled swir band to the 10m bands
    im_ms = np.append(im10, im_swir, axis=2)
    
    # read 60m band (QA)
    fn = os.path.join(filepath60, filenames60[i])
    data = gdal.Open(fn, gdal.GA_ReadOnly)
    bands = [data.GetRasterBand(k + 1).ReadAsArray() for k in range(data.RasterCount)]
    im60 = np.stack(bands, 2)
    im_qa = im60[:,:,0]
    cloud_mask = sds.create_cloud_mask(im_qa, satname, plot_bool)
    cloud_mask = transform.resize(cloud_mask,(nrows, ncols), order=0, preserve_range=True, mode='constant')
    # check if -inf or nan values on any band and add to cloud mask
    for k in range(im_ms.shape[2]):   
            im_inf = np.isin(im_ms[:,:,k], -np.inf)
            im_nan = np.isnan(im_ms[:,:,k])
            cloud_mask = np.logical_or(np.logical_or(cloud_mask, im_inf), im_nan)
            
    # calculate cloud cover and if above threshold, skip it
    cloud_cover = sum(sum(cloud_mask.astype(int)))/(cloud_mask.shape[0]*cloud_mask.shape[1])
    if cloud_cover > cloud_thresh:
        print('skip ' + str(i) + ' - cloudy (' + str(np.round(cloud_cover*100).astype(int)) + '%)')
        idx_skipped.append(i)
        continue
    
    # rescale image intensity for display purposes
    im_display = sds.rescale_image_intensity(im_ms[:,:,[2,1,0]], cloud_mask, 99.9, False)
    
    # plot rgb image
    plt.figure()
    plt.subplot(121)
    plt.axis('off')
    plt.imshow(im_display)
    
    # classify image in 4 classes (sand, whitewater, water, other) with NN classifier
    im_classif, im_labels = sds.classify_image_NN_nopan(im_ms, cloud_mask, min_beach_size, plot_bool)
    
    plt.subplot(122)
    plt.axis('off')
    im = np.copy(im_display)
    colours = np.array([[1,128/255,0/255],[204/255,1,1],[0,0,204/255]])
    for k in range(0,im_labels.shape[2]):
        im[im_labels[:,:,k],0] = colours[k,0]
        im[im_labels[:,:,k],1] = colours[k,1]
        im[im_labels[:,:,k],2] = colours[k,2]
    plt.imshow(im)
    # store the data
    cloud_cover_ts.append(cloud_cover)
    acc_georef_ts.append(metadata[satname]['acc_georef'][i])
    filename_ts.append(filenames10[i])
    satname_ts.append(satname)
    date_acquired_ts.append(filenames10[i][:10])
    timestamp.append(metadata[satname]['dates'][i])

# store in output structure
output[satname] = {'dates':timestamp, 'idx_skipped':idx_skipped,
      'metadata':{'filenames':filename_ts, 'satname':satname_ts, 'cloud_cover':cloud_cover_ts,
                  'acc_georef':acc_georef_ts}}

# save output
#with open(os.path.join(filepath, sitename + '_output' + satname + '.pkl'), 'wb') as f:
#    pickle.dump(output, f)