-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathanalysis.py
284 lines (235 loc) · 10.4 KB
/
analysis.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
import os
from pathlib import Path
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from scipy import stats
import shapely.geometry as shgeo
import time, shutil
# 设置风格，尺度
sns.set_style('darkgrid')
sns.set_context('paper')
# 支持中文
plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False # 用来正常显示负号
def TuplePoly2Poly(poly):
outpoly = [poly[0][0], poly[0][1],
poly[1][0], poly[1][1],
poly[2][0], poly[2][1],
poly[3][0], poly[3][1]
]
return outpoly
def get_data(label_dir, bbox_type="hbb"):
"""
data_dict =
{
'0'：{
'area':[],
'center':[],
'ratio':[]
}
}
"""
data_dict = {}
assert Path(label_dir).is_dir(), "label_dir is not exist"
txts = os.listdir(label_dir)
for txt in txts:
with open(os.path.join(label_dir, txt), 'r') as f:
lines = f.readlines()
for line in lines:
temp = line.split()
coor_list = list(map(lambda x: int(x), temp[1:]))
if bbox_type == "hbb":
# label 格式：class_num x1 y1 w h
area = coor_list[2] * coor_list[3]
center = (int(coor_list[0] + 0.5*coor_list[2]),
int(coor_list[1] + 0.5*coor_list[3]))
ratio = round(coor_list[2] / coor_list[3], 2)
elif bbox_type == "obb":
# label 格式：class_num x1 y1 x2 y2 x3 y3 x4 y4
ploy = [
(float(coor_list[0]), float(coor_list[1])),
(float(coor_list[2]), float(coor_list[3])),
(float(coor_list[4]), float(coor_list[5])),
(float(coor_list[6]), float(coor_list[7])),
]
gtpoly = shgeo.Polygon(ploy)
area = gtpoly.area
ploy_list = list(map(int, TuplePoly2Poly(ploy)))
xmin, ymin, xmax, ymax = min(ploy_list[0::2]), min(ploy_list[1::2]),\
max(ploy_list[0::2]), max(ploy_list[1::2])
width, height = xmax - xmin, ymax - ymin
center = [int(xmin + 0.5 * width), int(ymin + 0.5 * height)]
ratio = round(width / height, 2)
if temp[0] not in data_dict:
data_dict[temp[0]] = {}
data_dict[temp[0]]['area'] = []
data_dict[temp[0]]['center'] = []
data_dict[temp[0]]['ratio'] = []
data_dict[temp[0]]['area'].append(area)
data_dict[temp[0]]['center'].append(center)
data_dict[temp[0]]['ratio'].append(ratio)
return data_dict
def get_data2(label_dir, label_start_num, class_num, select_num):
"""[summary]
Args:
label_dir ([type]): 标签路径
label_start_num ([type]): label标注第一类开始数字0 or 1
class_num ([type]): 类别数量
select_num ([type]): 选择指定的类别数字
Returns:
[type]: [description]
"""
assert Path(label_dir).is_dir(), "label_dir is not exist"
txts = os.listdir(label_dir)
data_list =[]
for txt in txts:
cnt_list = [0] * class_num
temp = []
name = int(txt.split(".")[0])
temp.append(name)
with open(os.path.join(label_dir, txt), 'r') as f:
lines = f.readlines()
for line in lines:
label = int(line.split()[0])
if label_start_num == 1:
cnt_list[label - 1] += 1
elif label_start_num == 0:
cnt_list[label] += 1
temp.extend(cnt_list)
data_list.append(temp)
# 过滤出只含指定类别的文件名
data = np.array(data_list)
data = data[np.argsort(data[:,select_num])][::-1] # select_num列降序排序
index = np.where(data[...,select_num] > 0)
return data[index][:,0]
def copy_files(src_dir, det_dir, data_list):
for name in data_list:
img_path = os.path.join(src_dir, "images", str(name) + ".tif")
txt_path = os.path.join(src_dir, "labels", str(name) + ".txt")
new_img_path = os.path.join(det_dir, "images", str(name) + ".tif")
new_txt_path = os.path.join(det_dir, "labels", str(name) + ".txt")
shutil.copy(img_path, new_img_path)
shutil.copy(txt_path, new_txt_path)
print(name)
def analysis_bbox(data, label_path, save_name):
nums = len(data)
fig, axes = plt.subplots(3, nums, figsize=(10*nums, 3*5))
fig.suptitle("各类bbox统计图\n Label path : %s" % label_path, fontsize=20)
for i, class_num in enumerate(data):
# area
ax = sns.distplot(data[class_num]['area'],
norm_hist=False, kde=False, ax=axes[0][i])
ax.set_title("class: %s" % class_num)
ax.set_xlabel('area')
ax.set_ylabel('times')
area_arr = np.array(data[class_num]['area'])
nums_str = "nums : %d" % len(area_arr)
aver_str = "aver : %d" % np.mean(area_arr)
medi_str = "medi : %d" % np.median(area_arr)
mode_str = "mode : %d" % stats.mode(area_arr)[0][0]
ax.text(0.7, 0.9, nums_str, transform=ax.transAxes, color='red')
ax.text(0.7, 0.86, aver_str, transform=ax.transAxes, color='purple')
ax.text(0.7, 0.82, medi_str, transform=ax.transAxes, color='blue')
ax.text(0.7, 0.78, mode_str, transform=ax.transAxes, color='navy')
if i == 0 :
ax.text(-0.5,0.5, "bbox area\n 分布", transform=ax.transAxes, fontsize=20, color='blue')
# ratio
ax = sns.distplot(data[class_num]['ratio'],
norm_hist=False, kde=False, ax=axes[1][i])
ax.set_title("class: %s" % class_num)
ax.set_xlabel('ratio')
ax.set_ylabel('times')
ratio_arr = np.array(data[class_num]['ratio'])
nums_str = "nums : %d" % len(ratio_arr)
aver_str = "aver : %.2f" % np.mean(ratio_arr)
medi_str = "medi : %.2f" % np.median(ratio_arr)
mode_str = "mode : %.2f" % stats.mode(ratio_arr)[0][0]
ax.text(0.7, 0.9, nums_str, transform=ax.transAxes, color='red')
ax.text(0.7, 0.86, aver_str, transform=ax.transAxes, color='purple')
ax.text(0.7, 0.82, medi_str, transform=ax.transAxes, color='blue')
ax.text(0.7, 0.78, mode_str, transform=ax.transAxes, color='navy')
if i == 0 :
ax.text(-0.5,0.5, "bbox ratio\n 分布", transform=ax.transAxes, fontsize=20, color='blue')
# center
a = np.array(data[class_num]['center'])
#ax = sns.scatterplot(a[..., 0], a[..., 1], ax=axes[2][i], marker='D', s=30, edgecolors='red')
ax = sns.kdeplot(a[..., 0], a[..., 1], ax=axes[2][i], shade=True, bw="silverman", cmap="mako")
#ax = sns.kdeplot(a[..., 0], a[..., 1], ax=axes[2][i], fill=True, thresh=0.5, levels=100, cmap="mako")
ax.set_title("class: %s" % class_num)
if i == 0 :
ax.text(-0.5,0.5, "bbox center\n 分布", transform=ax.transAxes, fontsize=20, color='blue')
save_path = os.path.join("./AnalysisResults", save_name)
plt.savefig(save_path,dpi=500,bbox_inches = 'tight')
#plt.show()
def analysis_total(data, label_path, save_name):
x = []
y = []
for class_num in data:
x.append("class %s" % class_num)
y.append(len(data[class_num]['area']))
fig, axes = plt.subplots(1, 3, figsize=(40, 6))
fig.suptitle("数据集总体分析\n Label path : %s" % label_path, fontsize=10)
ax = sns.barplot(x, y, ax=axes[0])
ax.set_title("类别数量分布")
ax.set_xlabel('类别')
ax.set_ylabel('数量')
for i, v in enumerate(y, start=0):
ax.text(i, v, "%d(%.2f%%)" %
(v, v/sum(y) * 100), ha="center", va="bottom")
y = []
for class_num in data:
y.extend(data[class_num]['area'])
ax = sns.distplot(y, norm_hist=False, kde=False, ax=axes[1])
ax.set_title("anchor 面积分布")
ax.set_xlabel('area')
ax.set_ylabel('times')
area_arr = np.array(y)
nums_str = "nums : %d" % len(area_arr)
aver_str = "aver : %.2f" % np.mean(area_arr)
medi_str = "medi : %.2f" % np.median(area_arr)
mode_str = "mode : %.2f" % stats.mode(area_arr)[0][0]
max_str = "max : %.2f" % np.max(area_arr)
min_str = "min : %.2f" % np.min(area_arr)
ax.text(0.7, 0.9, nums_str, transform=ax.transAxes, color='red')
ax.text(0.7, 0.86, aver_str, transform=ax.transAxes, color='purple')
ax.text(0.7, 0.82, medi_str, transform=ax.transAxes, color='blue')
ax.text(0.7, 0.78, mode_str, transform=ax.transAxes, color='navy')
ax.text(0.7, 0.74, max_str, transform=ax.transAxes, color='coral')
ax.text(0.7, 0.70, min_str, transform=ax.transAxes, color='cyan')
y = []
for class_num in data:
y.extend(data[class_num]['ratio'])
ax = sns.distplot(y, norm_hist=False, kde=False, ax=axes[2])
ax.set_title("anchor ratio(W/H) 分布")
ax.set_xlabel('ratio')
ax.set_ylabel('times')
area_arr = np.array(y)
nums_str = "nums : %d" % len(area_arr)
aver_str = "aver : %.2f" % np.mean(area_arr)
medi_str = "medi : %.2f" % np.median(area_arr)
mode_str = "mode : %.2f" % stats.mode(area_arr)[0][0]
max_str = "max : %.2f" % np.max(area_arr)
min_str = "min : %.2f" % np.min(area_arr)
ax.text(0.7, 0.9, nums_str, transform=ax.transAxes, color='red')
ax.text(0.7, 0.86, aver_str, transform=ax.transAxes, color='purple')
ax.text(0.7, 0.82, medi_str, transform=ax.transAxes, color='blue')
ax.text(0.7, 0.78, mode_str, transform=ax.transAxes, color='navy')
ax.text(0.7, 0.74, max_str, transform=ax.transAxes, color='coral')
ax.text(0.7, 0.70, min_str, transform=ax.transAxes, color='cyan')
save_path = os.path.join("./AnalysisResults", save_name)
plt.savefig(save_path,dpi=500,bbox_inches = 'tight')
#plt.show()
def analysis(label_dir, bbox_type="hbb"):
data = get_data(label_dir, bbox_type)
if not os.path.exists("./AnalysisResults"):
os.makedirs("./AnalysisResults")
save_time = time.strftime('%Y-%m-%d_%H-%M-%S',time.localtime(time.time()))
analysis_bbox(data, label_dir, "bbox_"+ save_time + ".jpg")
analysis_total(data, label_dir, "total_"+ save_time + ".jpg")
plt.show()
if __name__ == "__main__":
label_dir = r"E:\研究所数据集\labels"
data = get_data2(label_dir, 1, 5, 3)
print(data)
copy_files(r"E:\研究所数据集", r"E:\研究所数据集\class3", data)