L8S2WaterDetector/AutomatedClassificationProcessor.py at master · ChrisLisbon/L8S2WaterDetector · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sun Jun 21 14:38:29 2020

@author: julia
"""

import os
import gdal
import numpy as np
from DataPreparationClass import DataPreparator
from IndicesCalculatorClass import IndicesCalculator
from WatershesBasedClassifierClass import WatershesBasedClassifier
from primary_functions import get_binary_classified_array, save_array_as_gtiff, get_binary_array_from_clasters, reverse_binary_array

class ClassificationProcessor:
    def __init__(self, input_directory, output_directory,
                 landsat_correction_method='srem',#dos
                 usgs_util_path= None,
                 landsat_cloud_fmask=False,
                 sentinel2_cloud=None, # fmask, s2cloudless, native_2A_level(https://earth.esa.int/web/sentinel/technical-guides/sentinel-2-msi/level-2a/algorithm)  C1 - sen2cor -
                 sentinel2_resolution=10, # 10 - with interpolation in B6 and B7, 20 - without interpolation
                 sen2cor_util_path=None,
                 landsat=False, sentinel1=False, sentinel2=False,
                 ):
        self.input_directory=input_directory
        self.output_directory=output_directory

        self.usgs_util_path=usgs_util_path
        self.sen2cor_util_path=sen2cor_util_path
        self.sentinel2_resolution=sentinel2_resolution
        self.landsat_correction_method=landsat_correction_method
        self.landsat_cloud_fmask=landsat_cloud_fmask
        self.sentinel2_cloud=sentinel2_cloud

        self.landsat=landsat
        self.sentinel1=sentinel1
        self.sentinel2=sentinel2

    def prepare_dataset(self, outputBounds=None, outputBoundsSRS=None): #outputBounds (minX, minY, maxX, maxY)
        prep=DataPreparator(self.input_directory,
                            landsat_correction_method=self.landsat_correction_method,
                             usgs_util_path=self.usgs_util_path,
                             landsat_cloud_fmask=self.landsat_cloud_fmask,
                             sentinel2_cloud=self.sentinel2_cloud,
                             sentinel2_resolution=self.sentinel2_resolution,
                             sen2cor_util_path=self.sen2cor_util_path
                            )
        prep.prepare_datasets(self.output_directory, landsat=self.landsat, sentinel1=self.sentinel1, sentinel2=self.sentinel2)
        prep=None
        if outputBounds!=None:
            for folder in os.listdir(self.output_directory):
                for file in os.listdir(os.path.join(self.output_directory, folder)):
                    full_file=os.path.join(self.output_directory, folder, file)
                    gdal.Warp(full_file, full_file, format = 'GTiff', outputBounds=outputBounds, outputBoundsSRS = outputBoundsSRS, dstSRS=outputBoundsSRS)

    def calculate_indices(self, landsat=False, sentinel2=False):
        for folder in os.listdir(self.output_directory):
            if folder=='landsat' and landsat==True:
                ind_cal=IndicesCalculator(os.path.join(self.output_directory, folder))
                os.mkdir(os.path.join(self.output_directory, 'landsat_indices'))
                ind_cal.save_indices(os.path.join(self.output_directory, 'landsat_indices'))

            if folder=='sentinel2' and sentinel2==True:
                ind_cal=IndicesCalculator(os.path.join(self.output_directory, folder))
                os.mkdir(os.path.join(self.output_directory, 'sentinel2_indices'))
                ind_cal.save_indices(os.path.join(self.output_directory, 'sentinel2_indices'))
        try:
            os.remove(self.output_directory+'/landsat/cloud_mask.tif.aux.xml')
        except Exception as e:
            print(e)
            pass

    def classify_dataset(self, landsat=False, sentinel2=False, bands_using=True, indices_using=True):
        if sentinel2==True:
            images_collection=[]
            if indices_using==True:
                indices=os.listdir(os.path.join(self.output_directory, 'sentinel2_indices'))
                for index in indices:
                    images_collection.append(os.path.join(self.output_directory, 'sentinel2_indices', index))
            if bands_using==True:
                bands=os.listdir(os.path.join(self.output_directory, 'sentinel2'))
                for band in bands:
                    if band!='cloud_mask.tif':
                        images_collection.append(os.path.join(self.output_directory, 'sentinel2', band))
            a=WatershesBasedClassifier(images_collection, base_image_index=2)
            a.get_classified_segmentation(os.path.join(self.output_directory, 'sentinel2_class.tif'), mode='raster', window_size=500, statistical_indicators=['mean', 'min', 'max'])
            a=None
            for index in indices:
                if index=='NDWI.tif':
                    ds=gdal.Open(os.path.join(self.output_directory, 'sentinel2_indices', index))
                    array=np.array(ds.GetRasterBand(1).ReadAsArray())
                    ds=None
                    ndwi_bin_array=get_binary_classified_array(array)
                    save_array_as_gtiff(ndwi_bin_array, os.path.join(self.output_directory, 'ndwi_bin_array.tif'), gtiff_path=os.path.join(self.output_directory, 'sentinel2_class.tif'))
                    array=None
                if 'NDVI' in index:
                    ds=gdal.Open(os.path.join(self.output_directory, 'sentinel2_indices', index))
                    array=np.array(ds.GetRasterBand(1).ReadAsArray())
                    ds=None
                    ndvi_bin_array=get_binary_classified_array(array)
                    save_array_as_gtiff(ndvi_bin_array, os.path.join(self.output_directory, 'ndvi_bin_array.tif'), gtiff_path=os.path.join(self.output_directory, 'sentinel2_class.tif'))
                    array=None
            claster_ds=gdal.Open(os.path.join(self.output_directory, 'sentinel2_class.tif'))
            claster_array=np.array(claster_ds.GetRasterBand(1).ReadAsArray())
            claster_ds=None
            new_claster_array=get_binary_array_from_clasters(claster_array, [ndwi_bin_array, reverse_binary_array(ndvi_bin_array)])

            mask_path=os.path.join(self.output_directory, 'sentinel2', 'nir.tif')
            mask_ds=gdal.Open(mask_path)
            mask_array=mask_ds.getRasterBand(1).ReadAsArray()
            mask_nodata=mask_ds.getRasterBand(1).GetNoDataValue()
            mask_ds=None
            new_claster_array[mask_array==mask_nodata]=5
            mask_array=None

            ndwi_bin_array=None
            ndvi_bin_array=None
            save_array_as_gtiff(new_claster_array, os.path.join(self.output_directory, 'sentinel2_water_mask.tif'), gtiff_path=os.path.join(self.output_directory, 'sentinel2_class.tif'), dtype='uint', nodata_value=5)

        if landsat==True:
            images_collection=[]
            if indices_using==True:
                indices=os.listdir(os.path.join(self.output_directory, 'landsat_indices'))
                for index in indices:
                    images_collection.append(os.path.join(self.output_directory, 'landsat_indices', index))
            if bands_using==True:
                bands=os.listdir(os.path.join(self.output_directory, 'landsat'))
                for band in bands:
                    if band!='cloud_mask.tif':
                        images_collection.append(os.path.join(self.output_directory, 'landsat', band))
            a=WatershesBasedClassifier(images_collection, base_image_index=2)
            a.get_classified_segmentation(os.path.join(self.output_directory, 'landsat_class.tif'), mode='raster', window_size=500, statistical_indicators=['mean', 'min', 'max'])
            a=None
            for index in indices:
                if index=='NDWI.tif':
                    ds=gdal.Open(os.path.join(self.output_directory, 'landsat_indices', index))
                    array=np.array(ds.GetRasterBand(1).ReadAsArray())
                    ds=None
                    ndwi_bin_array=get_binary_classified_array(array)
                    save_array_as_gtiff(ndwi_bin_array, os.path.join(self.output_directory, 'ndwi_bin_array.tif'), gtiff_path=os.path.join(self.output_directory, 'landsat_class.tif'), dtype='uint')
                    array=None
                if 'NDVI' in index:
                    ds=gdal.Open(os.path.join(self.output_directory, 'landsat_indices', index))
                    array=np.array(ds.GetRasterBand(1).ReadAsArray())
                    ds=None
                    ndvi_bin_array=get_binary_classified_array(array)
                    save_array_as_gtiff(ndvi_bin_array, os.path.join(self.output_directory, 'ndvi_bin_array.tif'), gtiff_path=os.path.join(self.output_directory, 'landsat_class.tif'), dtype='uint')
                    array=None
            claster_ds=gdal.Open(os.path.join(self.output_directory, 'landsat_class.tif'))
            claster_array=np.array(claster_ds.GetRasterBand(1).ReadAsArray())
            claster_ds=None
            new_claster_array=get_binary_array_from_clasters(claster_array, [ndwi_bin_array, reverse_binary_array(ndvi_bin_array)])

            mask_path=os.path.join(self.output_directory, 'landsat', 'nir.tif')
            mask_ds=gdal.Open(mask_path)
            mask_array=mask_ds.getRasterBand(1).ReadAsArray()
            mask_nodata=mask_ds.getRasterBand(1).GetNoDataValue()
            mask_ds=None
            new_claster_array[mask_array==mask_nodata]=5
            mask_array=None

            ndwi_bin_array=None
            ndvi_bin_array=None
            save_array_as_gtiff(new_claster_array, os.path.join(self.output_directory, 'landsat_water_mask.tif'), gtiff_path=os.path.join(self.output_directory, 'landsat_class.tif'), dtype='uint', nodata_value=5)

    def create_consolidated_water_mask(self):
        sentinel2_mask_path=os.path.join(self.output_directory, 'sentinel2_water_mask.tif')
        sentinel2_mask_ds=gdal.Open(sentinel2_mask_path)
        sentinel2_mask_nodata_value=sentinel2_mask_ds.GetRasterBand(1).GetNoDataValue()
        sentinel2_mask_array=sentinel2_mask_ds.GetRasterBand(1).ReadAsArray()
        sentinel2_mask_ds=None

        landsat_mask_path=os.path.join(self.output_directory, 'landsat_water_mask.tif')
        landsat_mask_ds=gdal.Warp('MEM', landsat_mask_path,  options=gdal.WarpOptions(format = 'MEM', xRes = 10, yRes = -10))
        landsat_mask_nodata_value=landsat_mask_ds.GetRasterBand(1).GetNoDataValue()
        landsat_mask_array=landsat_mask_ds.GetRasterBand(1).ReadAsArray()
        landsat_mask_ds=None

        sentinel2_mask_array[sentinel2_mask_array==sentinel2_mask_nodata_value]=landsat_mask_array

        save_array_as_gtiff(sentinel2_mask_array, os.path.join(self.output_directory, 'sentinel2_landsat_water_mask.tif'), gtiff_path=os.path.join(self.output_directory, 'sentinel2_water_mask.tif'), dtype='uint', nodata_value=landsat_mask_nodata_value)


output_folder='/media/julia/Data/water_detection_sbp/out'
input_folder='/media/julia/Data/water_detection_sbp'

a=ClassificationProcessor(input_folder, output_folder, sentinel2=True, landsat=True,
                          landsat_correction_method='dos',
                          landsat_cloud_fmask=True, sentinel2_cloud='s2cloudless')
#a.prepare_dataset(outputBounds=[609684.7559, 6639270.6362, 634680.5438, 6658914.9030], outputBoundsSRS='EPSG:32635')
#a.calculate_indices(sentinel2=True, landsat=True)
a.classify_dataset(sentinel2=True, landsat=True)