-
Notifications
You must be signed in to change notification settings - Fork 0
/
whiteboxtool_processing.py
327 lines (274 loc) · 13.2 KB
/
whiteboxtool_processing.py
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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
#%%
from whitebox_tools import WhiteboxTools
# from import_export_geotiff import ImportGeoTiff, ExportGeoTiff
import tqdm
# from skimage.feature import peak_local_max
import rasterio as rio
import numpy as np
from scipy import ndimage as ndi
import os
import rioxarray as rxr
import matplotlib.pyplot as plt
from matplotlib import colors
import pandas as pd
import seaborn as sns
from rasterio.enums import Resampling
import geopandas as gp
import datetime
import richdem as rd
import matplotlib
def whiteboxtool_calculations(directory, first_date, ref_date, mask):
#for testing:
# directory = "/home/torka/PraktikumAWI/"
# mask = "leg4_icefloe_leftCutout.shp"
# first_date = "20200423"
# ref_date = "20200630"
#----------
os.chdir(directory)
DEMname = first_date #+ "_01_DEM_int_PS_crop_05m_shift_crop_to_shape_filledNan_crop_to_shape.tif"
class_file = ref_date + '_01_main_classes_crop_to_shape.tif'
dem = rd.LoadGDAL(DEMname)
os.chdir(directory)
wbt = WhiteboxTools()
wbt.work_dir = directory
flooded_dem = wbt.fill_depressions(DEMname, DEMname[:-4] + '_wbt_flooded_dem.tif', fix_flats=False);
flooded_dem = rd.LoadGDAL(DEMname[:-4] + '_wbt_flooded_dem.tif')
# depths_tif = ExportGeoTiff(DEMname[0:8] + "_wbt_depth.tif",depth,width,height,match_geotrans,match_proj)
# depths_tif = ExportGeoTiff(DEMname[0:8] + "_wbt_depth.tif",depth,width,height,match_geotrans,match_proj) #second time, else the function write an empty .tif
# accum_d8 = wbt.d8_flow_accumulation(DEMname, DEMname[:-4] + '_wbt_d8_flowacc.tif');
os.chdir(directory)
depth = flooded_dem - dem
depth[depth == 0] = np.nan
print('depths determined...')
########### terrain approach statistics and plot
rxr_clas = rxr.open_rasterio(class_file)
rxr_depths = rxr.open_rasterio(DEMname)
floe_shape=gp.read_file(mask)
clas=rxr_clas.rio.clip(floe_shape.geometry).values[0,:,:]
# clas = np.where(clas == 255, np.nan, np.where(clas == 0, True, False))
raster_tmp = np.invert(np.isnan(rxr_depths))
raster = rxr_depths.where(raster_tmp,0)
hydr_tmp=raster.rio.clip(floe_shape.geometry).values[0,:,:]
clas = np.where(clas == 255, np.nan, np.where(clas == 0, 0, 1))
clas = np.rint(clas)
hydr_tmp = np.isnan(hydr_tmp)
hydr_tmp = np.where(hydr_tmp,np.nan,1)
depth = np.where(depth>0.03,0,1)
hydr = np.where(hydr_tmp==1,depth,hydr_tmp)
#hydr plot
# col_dict_hydr={0:"dodgerblue",1:"whitesmoke"}
# cm_hydr = colors.ListedColormap([col_dict_hydr[x] for x in col_dict_hydr.keys()])
# labels_hydr = np.array(["meltpond","no meltpond"])
# len_lab_hydr = len(labels_hydr)
# norm_bins_hydr = np.sort([*col_dict_hydr.keys()]) + 0.5
# norm_bins_hydr = np.insert(norm_bins_hydr, 0, np.min(norm_bins_hydr) - 1.0)
# norm_hydr = matplotlib.colors.BoundaryNorm(norm_bins_hydr, len_lab_hydr, clip=True)
# fmt_hydr = matplotlib.ticker.FuncFormatter(lambda x, pos: labels_hydr[norm_hydr(x)])
# fig,ax= plt.subplots()
# plot = ax.imshow(hydr,cmap=cm_hydr,norm=norm_hydr)
# diff_hydr = norm_bins_hydr[1:] - norm_bins_hydr[:-1]
# tickz_hydr = norm_bins_hydr[:-1] + diff_hydr / 2
# # cb_hydr = fig.colorbar(plot, format=fmt_hydr, ticks=tickz_hydr,ax=ax,location='bottom',shrink=0.4)
# ax.axis('off')
# plt.tight_layout()
# plt.show()
#preparing and exporting an accuracy map
accuracy_map = np.empty((len(clas),len(clas[0])))
accuracy_map[:] = -1
TP=np.sum(np.logical_and(clas==0, hydr==0))
FP=np.sum(np.logical_and(clas==1, hydr==0))
FN=np.sum(np.logical_and(clas==0, hydr==1))
TN=np.sum(np.logical_and(clas==1, hydr==1)) #subtract all pixels outside of clipped area #subtract all pixels outside of clipped area
precision=(TP/float(TP+FP))*100
recall=(TP/float(TP+FN))*100
accuracy=((TP+TN)/float(TP+FP+FN+TN))*100
bal_accuracy=((TP/float(TP+TN)+TN/float(TN+FP))/2)*100
fscore= 2*np.array(precision)*np.array(recall)/(np.array(precision)+np.array(recall))
# for i in tqdm.tqdm(range(0,len(clas))):
# for j in range(0,len(clas[0])):
# if clas[i,j] == 1 and hydr[i,j] == 1:
# accuracy_map[i,j] = 0 #TN
# elif clas[i,j] == 1 and hydr[i,j] == 0:
# accuracy_map[i,j] = 1 #FP
# elif clas[i,j] == 0 and hydr[i,j] == 1:
# accuracy_map[i,j] = 2 #FN
# elif clas[i,j] == 0 and hydr[i,j] == 0:
# accuracy_map[i,j] = 3 #TP
# accuracy_map = np.where(accuracy_map == -1, np.nan, accuracy_map)
# col_dict={0:"mediumseagreen",1:"indianred",2:"gold",3:"dodgerblue"}
# cm = colors.ListedColormap([col_dict[x] for x in col_dict.keys()])
# labels = np.array(["TN","FP","FN","TP"])
# len_lab = len(labels)
# norm_bins = np.sort([*col_dict.keys()]) + 0.5
# norm_bins = np.insert(norm_bins, 0, np.min(norm_bins) - 1.0)
# print(norm_bins)
# norm = matplotlib.colors.BoundaryNorm(norm_bins, len_lab, clip=True)
# fmt = matplotlib.ticker.FuncFormatter(lambda x, pos: labels[norm(x)])
# fig, axs =plt.subplots(1,2,figsize=(10,5), gridspec_kw={'width_ratios': [2, 1]})
# # fig.suptitle("Pysheds confusion \n"+ first_date + " vs. " + ref_date)
# im = axs[0].imshow(accuracy_map, cmap=cm, norm=norm)
# diff = norm_bins[1:] - norm_bins[:-1]
# tickz = norm_bins[:-1] + diff / 2
# cb = fig.colorbar(im, format=fmt, ticks=tickz,ax=axs[0],location='left',shrink=0.8)
# axs[0].text(-0.18,1,'a',ha='left',va='center',transform=axs[0].transAxes,fontsize=14,weight="bold")
# axs[0].axis("off")
# #accuracy_tif = ExportGeoTiff(first_date + 'vs' + ref_date + '_RichDEM_accuracy.tif',accuracy_map,width,height,match_geotrans,match_proj)
# #accuracy_tif = ExportGeoTiff(first_date + 'vs' + ref_date + '_RichDEM_accuracy.tif',accuracy_map,width,height,match_geotrans,match_proj)
# data = {'y_Actual':clas.flatten("F"),'y_Predicted':hydr.flatten("F")}
# df = pd.DataFrame(data, columns=['y_Actual','y_Predicted'])
# df = df.dropna()
# df = df[(df["y_Actual"] != 2) & (df["y_Actual"] != 3)]
# cf_matrix = pd.crosstab(df['y_Actual'], df['y_Predicted'], rownames=['Actual'], colnames=['Predicted'], normalize='index')
# axs[1]= sns.heatmap(cf_matrix, annot=True, cmap='Blues',square=True,cbar=False)
# axs[1].text(-0.1,1.1,'b',ha='left',va='center',transform=axs[1].transAxes,fontsize=14,weight="bold")
# #title = 'Confusion Matrix - RichDEM \n ' + first_date + ' vs ' + ref_date
# #ax.set_title(title);
# axs[1].set_xlabel('\nPredicted Values')
# axs[1].set_ylabel('Actual Values ');
# ## Ticket labels - List must be in alphabetical order
# axs[1].xaxis.set_ticklabels(['True','False'])
# axs[1].yaxis.set_ticklabels(['True','False'])
# ## Display the visualization of the Confusion Matrix.
# plt.show()
return(precision,recall,accuracy,fscore,bal_accuracy)
#user inputs:
#Load DEM
first_date = ['20200107_01_ALS_DEM_05m_leg4CO_shift_filledNan_crop_to_shape.tif','20200116_01_ALS_DEM_05m_leg4CO_shift_filledNan_crop_to_shape.tif','20200321_ALS_DEM_filledNan_crop_to_shape.tif','20200321_01_DEM_int_PS_crop_05m_shift_crop_to_shape_filledNan_crop_to_shape.tif','20200423_01_ALS_DEM_05m_leg4CO_shift_crop_to_shape.tif','20200423_01_DEM_int_PS_crop_05m_shift_crop_to_shape_filledNan_crop_to_shape.tif'] #20200321, 20200423, 20200510 are valid options
ref_date = ['20200630', '20200707', '20200717', '20200722'] #20200630, 20200704, 20200707, 20200717, 20200722 are valid options
directory = "/home/torka/PraktikumAWI/"
mask = "leg4_icefloe_leftCutout.shp"
# format_first_date = [datetime.datetime.strptime(i, '%Y%m%d') for i in first_date]
# str_first_date = [i.strftime('%d.%m.%Y') for i in format_first_date]
str_first_date = ['07.01\nALS','16.01\nALS','21.03\nALS','21.03\nphoto','23.04\nALS','23.04\nphoto']
format_ref_date = [datetime.datetime.strptime(i, '%Y%m%d') for i in ref_date]
str_ref_date = [i.strftime('%d.%m.%Y') for i in format_ref_date]
precision_list = []
recall_list = []
accuracy_list = []
fscore_list = []
balacc_list = []
d_optimal_list = []
for ref in ref_date:
for first in first_date:
precision,recall,accuracy,fscore,bal_accuracy = whiteboxtool_calculations(directory, first, ref, mask)
precision_list.append(precision)
recall_list.append(recall)
accuracy_list.append(accuracy)
fscore_list.append(fscore)
balacc_list.append(bal_accuracy)
#%%
#%%
shape = (len(ref_date),len(first_date))
precision_array = np.array(precision_list)
precision_array = precision_array.reshape(shape)
ax = sns.heatmap(precision_array, annot=True, cmap='magma_r',vmin=25,vmax=50,fmt=".1f",annot_kws={"size": 12})
# ax.set_title('Precision Matrix - wbt \n with d=0.13' );
ax.set_xlabel('Prediction dates')
ax.set_ylabel('Reference dates');
## Ticket labels - List must be in alphabetical order
ax.xaxis.set_ticklabels(str_first_date)
ax.yaxis.set_ticklabels(str_ref_date)
ax.invert_yaxis()
## Display the visualization of the Confusion Matrix.
plt.show()
#%%
shape = (len(ref_date),len(first_date))
recall_array = np.array(recall_list)
recall_array = recall_array.reshape(shape)
ax = sns.heatmap(recall_array, annot=True, cmap='magma_r',vmin=65,vmax=95,fmt=".1f",annot_kws={"size": 12})
# ax.set_title('Recall Matrix - wbt \n with d=0.13');
ax.set_xlabel('Prediction dates')
ax.set_ylabel('Reference dates');
## Ticket labels - List must be in alphabetical order
ax.xaxis.set_ticklabels(str_first_date)
ax.yaxis.set_ticklabels(str_ref_date)
ax.invert_yaxis()
## Display the visualization of the Confusion Matrix.
plt.show()
#%%
# shape = (len(ref_date),len(first_date))
# accuracy_array = np.array(accuracy_list)
# accuracy_array = accuracy_array.reshape(shape)
# ax = sns.heatmap(accuracy_array, annot=True, cmap='magma_r',robust=True,fmt=".1f",annot_kws={"size": 12})
# ax.set_title('Accuracy Matrix - wbt \n with d=0.13');
# ax.set_xlabel('Prediction dates')
# ax.set_ylabel('Reference dates');
# ## Ticket labels - List must be in alphabetical order
# ax.xaxis.set_ticklabels(str_first_date)
# ax.yaxis.set_ticklabels(str_ref_date)
# ax.invert_yaxis()
# ## Display the visualization of the Confusion Matrix.
# plt.show()
#%%
shape = (len(ref_date),len(first_date))
fscore_array = np.array(fscore_list)
fscore_array = fscore_array.reshape(shape)
ax = sns.heatmap(fscore_array, annot=True, cmap='magma_r',vmax=60,vmin=42,fmt=".1f",annot_kws={"size": 12})
# ax.set_title('F-score Matrix - wbt \n with d=0.13');
ax.set_xlabel('Prediction dates')
ax.set_ylabel('Reference dates');
## Ticket labels - List must be in alphabetical order
ax.xaxis.set_ticklabels(str_first_date)
ax.yaxis.set_ticklabels(str_ref_date)
ax.invert_yaxis()
## Display the visualization of the Confusion Matrix.
plt.show()
# plt.savefig("fscore_wbt.png")
#%%
shape = (len(ref_date),len(first_date))
balacc_array = np.array(balacc_list)
balacc_array = balacc_array.reshape(shape)
ax = sns.heatmap(balacc_array, annot=True, cmap='magma_r',robust=True,fmt=".1f",annot_kws={"size": 12})
# ax.set_title('Balanced Accuracy Matrix - wbt \n with d=0.13');
ax.set_xlabel('Prediction dates')
ax.set_ylabel('Reference dates');
## Ticket labels - List must be in alphabetical order
ax.xaxis.set_ticklabels(first_date)
ax.yaxis.set_ticklabels(ref_date)
ax.invert_yaxis()
## Display the visualization of the Confusion Matrix.
plt.show()
#%%
######### accumulation approach statistics and plot
# precision=[]
# recall=[]
# accuracy=[]
# depth_list=[]
# rxr_map = rxr.open_rasterio(DEMname) # use DEM file for georeference metadata, fill with pond data later
# rxr_clas = rxr.open_rasterio(class_file)
# # same as before
# rxr_map.values[0,:,:]=pond_vol_bool
# rxr_map_reproj = rxr_map.rio.reproject_match(rxr_clas,Resampling.nearest)
# clas=rxr_clas.rio.clip(floe_shape.geometry).values[0,:,:]
# clas=clas==0
# clas_back = rxr_clas.rio.clip(floe_shape.geometry).values[0,:,:]
# rxr_map=rxr_map_reproj.rio.clip(floe_shape.geometry).values[0,:,:]
# hydr=rxr_flooded>0.05 # let's only consider areas deeper than 5cm
# TP_half=np.sum(np.logical_and(clas, rxr_map==1)) # not fully filled basins
# TP_full=np.sum(np.logical_and(clas, rxr_map==2)) # overflowing basins
# TP_both=np.sum(np.logical_and(clas, np.isin(rxr_map,[1,2])))
# TP_05=np.sum(np.logical_and(clas, hydr)) # 5cm threshold
# FP_05=np.sum(np.logical_and(hydr, clas==False))
# FP_half=np.sum(np.logical_and(clas==False, rxr_map==1))
# FP_full=np.sum(np.logical_and(clas==False, rxr_map==2))
# FP_both=np.sum(np.logical_and(clas==False, np.isin(rxr_map,[1,2])))
# precision_half=TP_half/float(TP_half+FP_half)
# precision_full=TP_full/float(TP_full+FP_full)
# precision_both=TP_both/float(TP_both+FP_both)
# precision_05=TP_05/float(TP_05+FP_05)
# plot accumulation volume map
# fig, ax = plt.subplots(figsize=(8,6))
# fig.patch.set_alpha(0)
# plt.grid('on', zorder=0)
# im = ax.imshow(accum_d8_arr*volume, zorder=2,
# cmap='cubehelix',
# norm=colors.LogNorm(volume, (accum_d8_arr*volume).max()),
# interpolation='bilinear')
# cbar=plt.colorbar(im, ax=ax)
# plt.title('Flow Accumulation', size=14)
# plt.xlabel('Longitude')
# plt.ylabel('Latitude')
# plt.tight_layout()
# cbar.ax.tick_params(labelsize=16)
# cbar.set_label(label=r'Upstream volume [m$^3$]', size=18)
# plt.show()