histogram_equalize/histogram_equalize.py at master · rupestre-campos/histogram_equalize · 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
from osgeo import gdal
from osgeo.gdalconst import *
from collections import Counter
import numpy as np
from subprocess import call
import sys
import time
from collections import defaultdict

def dsum(*dicts):
    ret = defaultdict(int)
    for d in dicts:
        for k, v in d.items():
            ret[k] += v
    return dict(ret)

def read_histograms(n_bands,in_ds,cols,rows):
    freq_tot = {}
    for b in range(n_bands):
        b+=1
        band = in_ds.GetRasterBand(b)
        rasterArray = band.ReadAsArray(0,0,cols,rows)
        unique, counts = np.unique(rasterArray.flatten(), return_counts=True)
        freq = dict(zip(unique,counts))
        #freq = Counter(rasterArray.flatten())
        freq_tot = dsum(freq_tot,freq)
    return freq_tot

def equalize_histogram(img,img_type,in_nodata,out_nodata):
    call('gdal_edit -a_nodata {} {}'.format(in_nodata,img),shell=True)
    if img_type == 8:
        scale = 255
    elif img_type == 16:
        scale = 65535
    elif img_type == 32:
        scale = 1
    in_ds = gdal.Open(img)
    n_bands = in_ds.RasterCount
    driver = in_ds.GetDriver()
    rows = in_ds.RasterYSize
    cols = in_ds.RasterXSize
    out_path = '{}_hist{}'.format(img[:-4],img[-4:])
    if img_type != 32:
        if img_type == 16:
            out_ds = driver.Create(out_path, cols, rows, n_bands, GDT_Int16)
        elif img_type == 8:
            out_ds = driver.Create(out_path, cols, rows, n_bands, GDT_Byte)
    else:
        out_ds = driver.Create(out_path, cols, rows, n_bands, GDT_Float32)
    pdf = defaultdict(int)
    print("computing pdf...")
    freq_tot = read_histograms(n_bands,in_ds,cols,rows)

    size = cols*rows*n_bands - freq_tot[in_nodata]
    for val in freq_tot:
        pdf[val] += float(freq_tot[val])/size
    print("computing cdf...")
    value_list = sorted([i for i in pdf])
    value_list.remove(in_nodata)
    cdf = {}
    summing = 0
    for val in value_list:
        summing += pdf[val]
        new_val = summing*scale
        if new_val == 0:
            new_val = 1
        cdf[val] = new_val
    print("converting band histograms...")
    for b in range(n_bands):
        b+=1
        band = in_ds.GetRasterBand(b)
        raster_array = band.ReadAsArray(0,0,cols,rows)
        if img_type != 32:
            if img_type == 16:
                out_data = np.zeros((rows,cols), np.uint16)
            elif img_type == 8:
                out_data = np.zeros((rows,cols), np.uint8)
        else:
            out_data = np.zeros((rows,cols), np.float32)
        out_data = write_data(raster_array,out_data,cdf,in_nodata,rows,cols)
        out_band = out_ds.GetRasterBand(b)
        out_band.WriteArray(out_data, 0, 0)
        out_band.SetNoDataValue(out_nodata)
        out_band.FlushCache()
        out_band = out_data = None
    out_ds.SetGeoTransform(in_ds.GetGeoTransform())
    out_ds.SetProjection(in_ds.GetProjection())
    raster_array = out_ds = in_ds = None
    return out_path

def write_data(raster_array,out_data,cdf,in_nodata,rows,cols):
    for i in range(0, rows):
        for j in range(0, cols):
            cell_value = raster_array[i,j]
            if cell_value != in_nodata:
                out_data[i,j] = cdf[cell_value]
    return out_data

def time_exec(time1,time2):
    time = time2 - time1
    if time < 60:
        sys.stdout.write("\tTime: {:.2f} seconds\n".format(time))
    elif time < 3600:
        time = time/60
        sys.stdout.write("\tTime: {:.2f} minutes\n".format(time))
    elif time < 86400:
        time = (time/60)/60
        sys.stdout.write("\tTime: {:.2f} hours\n".format(time))
    else:
        time = ((time/60)/60)/24
        sys.stdout.write("Time: {} days\n".format(time))

def main():
    img = sys.argv[1]
    out_img_type = int(sys.argv[2])
    in_nodata = int(sys.argv[3])
    out_nodata = int(sys.argv[4])
    time1 = time.time()
    print("starting to equalize...")
    result = equalize_histogram(img,out_img_type,in_nodata,out_nodata)
    time2 = time.time()
    sys.stdout.write("image equalized:{} \n".format(result))
    time_exec(time1,time2)

if __name__ == "__main__":
    main()