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
| import json
import cv2
import numpy as np
from matplotlib import pyplot as plt
import random
import os
import pandas as pd
import shutil
def cut(image):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(gray, 20, 255, cv2.THRESH_BINARY)
contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
max_contour = max(contours, key=cv2.contourArea)
x, y, w, h = cv2.boundingRect(max_contour)
cut = image[y:y+h,x:x+w]
return cut,[x,y,w,h]
def rotate(image,angle):
center = (image.shape[1]/2, image.shape[0]/2)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(gray, 20, 255, cv2.THRESH_BINARY)
contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
max_contour = max(contours, key=cv2.contourArea)
x, y, w, h = cv2.boundingRect(max_contour)
origin_corner_position = [[0, 0], [0, h], [w, 0], [h, w]]
rotation_matrix = cv2.getRotationMatrix2D(center, angle, 1)
for index, point in enumerate(origin_corner_position):
origin_corner_position[index] = np.matmul(rotation_matrix, np.array([point[0], point[1], 1]).T)
boundingBox = cv2.boundingRect(np.array(origin_corner_position, dtype=np.int32))
rotation_matrix[0][2] += (boundingBox[2] - w)/2
rotation_matrix[1][2] += (boundingBox[3] - h) / 2
rotated_image = cv2.warpAffine(image, rotation_matrix, (boundingBox[2], boundingBox[3]))
gray = cv2.cvtColor(rotated_image, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(gray, 20, 255, cv2.THRESH_BINARY)
contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
max_contour = max(contours, key=cv2.contourArea)
x, y, w, h = cv2.boundingRect(max_contour)
return rotated_image,(x,y,w,h)
def insert(bkg,img):
x = random.randint(0,bkg.shape[1]-img.shape[1])
y = random.randint(0,bkg.shape[0]-img.shape[0])
w = img.shape[1]
h = img.shape[0]
mb = img[:,:,0]
mg = img[:,:,1]
mr = img[:,:,2]
img[np.bitwise_and(np.bitwise_and(mb<30, mg<30), mr<30)]=0
bkg[y:y+h,x:x+w][img!=0]=0
bkg[y:y+h,x:x+w]+=img
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(gray, 20, 255, cv2.THRESH_BINARY)
contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
max_contour = max(contours, key=cv2.contourArea)
xx, yy, w, h = cv2.boundingRect(max_contour)
xx+=x
yy+=y
return bkg,(xx,yy,w,h)
class Item():
def __init__(self,label,img,pts):
self.label=label
pts=np.array(pts,int)
mask = np.zeros(img.shape)
mask=cv2.fillPoly(mask, [pts], (255, 255, 255))
img[mask==0]=0
self.cut=cut(img)[0]
def get(self):
image,_=rotate(self.cut,random.randint(0,180))
return image
def delete_folder(path):
try:
shutil.rmtree(path)
print(f"成功删除文件夹 {path}")
except OSError as e:
print(f"删除文件夹 {path} 失败: {e}")
def loadLabel(f):
with open(f, 'r', encoding='utf-8') as file:
data = json.load(file)
return data
def checkIOU(bboxes, threshold=0.5):
for i in range(len(bboxes)):
bbox_i = bboxes[i]
for j in range(i + 1, len(bboxes)):
bbox_j = bboxes[j]
x1_i, y1_i, w_i, h_i = bbox_i
x2_i, y2_i = x1_i + w_i, y1_i + h_i
x1_j, y1_j, w_j, h_j = bbox_j
x2_j, y2_j = x1_j + w_j, y1_j + h_j
x1_intersection = max(x1_i, x1_j)
y1_intersection = max(y1_i, y1_j)
x2_intersection = min(x2_i, x2_j)
y2_intersection = min(y2_i, y2_j)
minArea=min(w_i*h_i,w_j*h_j)
intersection_area = max(0, x2_intersection - x1_intersection) * max(0, y2_intersection - y1_intersection)
iou = intersection_area/minArea
if iou > threshold:
return True
return False
def getNames(everythings):
return [i.label for i in everythings]
def saveLabel(bboxs,types,filename):
inttype = [dics[i] for i in types]
bboxdf = pd.DataFrame(bboxs,columns=["x","y","w","h"])
bboxdf["cx"]=bboxdf["x"]+(bboxdf["w"]/2)
bboxdf["cy"]=bboxdf["y"]+(bboxdf["h"]/2)
bboxdf["cx"]/=bkg.shape[1]
bboxdf["w"]/=bkg.shape[1]
bboxdf["cy"]/=bkg.shape[0]
bboxdf["h"]/=bkg.shape[0]
bboxdf["t"]=inttype
bboxdf.loc[:,["t","cx","cy","w","h"],]
with open("{}.txt".format(filename),"w+") as f:
for i in np.array(bboxdf.loc[:,["t","cx","cy","w","h"],]):
f.write("{:d} {} {} {} {}\n".format(int(i[0]),i[1],i[2],i[3],i[4]))
def resize_image(image, max_size):
height, width = image.shape[:2]
if height <= max_size and width <= max_size:
return image
if height > width:
scale = max_size / height
else:
scale = max_size / width
new_height = int(height * scale)
new_width = int(width * scale)
resized_image = cv2.resize(image, (new_width, new_height))
return resized_image
def chooseTypeIndices(everythings,dic):
t=random.choice(list(dic.keys()))
return [i for i, item in enumerate(getNames(everythings)) if item == t ]
def chooseEqually(everythings,dic):
idx = chooseTypeIndices(everythings,dic)
return random.choice(idx)
|
