cover_classification/extract_cover_data.py at master · kdchadwick/cover_classification · 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


import gdal
import numpy as np
from tqdm import tqdm
import argparse
import pandas as pd
import os
import subprocess
import glob


def main():
    parser = argparse.ArgumentParser(
        description='efficiently extract data from a vector file and multiple accompanying rasters')

    parser.add_argument('out_base', type=str)
    parser.add_argument('-all_shape_dir',type=str)
    parser.add_argument('-shape_dirs', nargs='+', type=str)
    parser.add_argument('-shp_attribute', type=str, default='id')
    parser.add_argument('-max_samples_per_class', type=int, default=20000)
    parser.add_argument('-source_files', nargs='+', type=str)
    args = parser.parse_args()

    if args.all_shape_dir is not None:
        args.shape_dirs = glob.glob(os.path.join(args.all_shape_dir,'*'))

    if args.max_samples_per_class == -1:
        args.max_samples_per_class=1e15

    # Open / check all raster files.  Check is very cursory.
    file_sets = [gdal.Open(fi, gdal.GA_ReadOnly) for fi in args.source_files]
    n_features = 0
    for _f in range(len(file_sets)):
        assert file_sets[_f] is not None, 'Invalid input file'
        if (file_sets[_f].RasterXSize != file_sets[0].RasterXSize):
            print('Raster X Size does not match, terminiating')
            quit()
        if (file_sets[_f].RasterYSize != file_sets[0].RasterYSize):
            print('Raster Y Size does not match, terminiating')
            quit()
        n_features += file_sets[_f].RasterCount

    trans = file_sets[0].GetGeoTransform()
    namelist = []
    init=' -te {} {} {} {} -tr {} {} -init -1 '.format(
                          trans[0],
                          trans[3]+trans[5]*file_sets[0].RasterYSize,
                          trans[0]+trans[1]*file_sets[0].RasterXSize,
                          trans[3],
                          trans[1],
                          trans[5])

    cover_raster_file = os.path.join(args.out_base,'cover_raster.tif')
    if (os.path.isfile(cover_raster_file)):
        print('cover raster file already exists at {}, using'.format(cover_raster_file))
        for index, shape_dir in enumerate(args.shape_dirs):
            shape_files = glob.glob(shape_dir + '/*.geojson')
            dirnames = [x for x in shape_dir.split('/') if x != '']
            namelist.append(dirnames[-1])
    else:
        for index, shape_dir in enumerate(args.shape_dirs):

            shape_files = glob.glob(shape_dir + '/*.geojson')
            dirnames = [x for x in shape_dir.split('/') if x != '']
            namelist.append(dirnames[-1])

            for shpfile in shape_files:
                cmd_str = 'gdal_rasterize {} {} -burn {} {}'.format(
                          shpfile,
                          cover_raster_file,
                          index,
                          init,
                          )
                print(cmd_str)
                subprocess.call(cmd_str, shell=True)
                init=''

    # Open binary cover file
    # cover_set = gdal.Open(cover_raster_file, gdal.GA_ReadOnly)
    # cover_trans = cover_set.GetGeoTransform()
    # assert cover_set is not None, 'Invalid input file'
    #
    # # Get cover coordinates
    # covers = cover_set.ReadAsArray()
    # un_covers = np.unique(covers[covers != -1]).astype(int)
    #
    # coord_lists = []
    # num_outputs = 0
    # np.random.seed(13)
    # for cover in un_covers:
    #
    #     cover_coords = list(np.where(covers == cover))
    #     if len(cover_coords[0]) > args.max_samples_per_class:
    #         perm = np.random.permutation(len(cover_coords[0]))[:args.max_samples_per_class]
    #         cover_coords[0] = cover_coords[0][perm]
    #         cover_coords[1] = cover_coords[1][perm]
    #
    #     coord_lists.append(cover_coords)
    #     num_outputs += len(cover_coords[0])
    #
    #
    # # Read through files and grab relevant data
    # output_array = np.zeros((num_outputs, n_features + 3))
    # output_names = []
    #
    # start_index = 0
    # for cover in un_covers:
    #
    #     cover_coords = coord_lists[cover]
    #     for _line in tqdm(range(len(cover_coords[0])), ncols=80):
    #
    #         output_array[start_index + _line, 0] = covers[cover_coords[0][_line], cover_coords[1][_line]]
    #         output_array[start_index + _line, 1] = cover_coords[1][_line]*cover_trans[1]+cover_trans[0]
    #         output_array[start_index + _line, 2] = cover_coords[0][_line]*cover_trans[5]+cover_trans[3]
    #
    #         output_names.append(namelist[cover])
    #
    #         feat_ind = 3
    #         for _f in range(len(file_sets)):
    #             line = file_sets[_f].ReadAsArray(
    #                 0, int(cover_coords[0][_line]), file_sets[_f].RasterXSize, 1)
    #             if (len(line.shape) == 2):
    #                 line = np.reshape(line, (1, line.shape[0], line.shape[1]))
    #
    #             line = np.squeeze(line[..., cover_coords[1][_line]])
    #
    #             output_array[start_index + _line, feat_ind:feat_ind+file_sets[_f].RasterCount] = line.copy()
    #             feat_ind += file_sets[_f].RasterCount
    #
    #     start_index += len(cover_coords[0])
    #
    # output_names = np.array(output_names)
    #
    # # Export
    # header = ['ID', 'X_UTM', 'Y_UTM',]
    # for _f in range(len(file_sets)):
    #     header.extend([os.path.splitext(os.path.basename(args.source_files[_f]))[0]
    #                    [-4:] + '_B_' + str(n+1) for n in range(file_sets[_f].RasterCount)])
    # out_df = pd.DataFrame(data=output_array, columns=header)
    # out_df['covertype'] = output_names
    # out_df.to_csv(os.path.join(args.out_base,'cover_extraction.csv'),sep=',', index=False)


if __name__ == "__main__":
    main()