-
Notifications
You must be signed in to change notification settings - Fork 0
/
augmentation.py
358 lines (276 loc) · 12.1 KB
/
augmentation.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
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
from ultralytics import YOLO
import cv2
from PIL import Image
import random
from shapely.geometry import Polygon
import math
import numpy as np
from matplotlib.path import Path
import os
model = YOLO('yolov8l-seg.pt')
def pointGenCrop(ratio,img):
originalHeight=img.shape[0]
originalWidth=img.shape[1]
originalSize=[img.shape[1],img.shape[0]]
cropWidth=int(ratio*originalWidth)
cropHeight=int(ratio*originalHeight)
bottomRight=(originalWidth-cropWidth,originalHeight-cropHeight)
#print(bottomRight)
start_x = random.randint(0, bottomRight[0])
start_y = random.randint(0, bottomRight[1])
end_x = start_x + cropWidth
end_y = start_y + cropHeight
#print("start", start_x, start_y,"crop",cropWidth,cropHeight, "end", end_x, end_y)
p1 = (start_x / originalWidth, start_y / originalHeight)
p2 = (end_x / originalWidth, start_y / originalHeight)
p3 = (end_x / originalWidth, end_y / originalHeight)
p4 = (start_x / originalWidth, end_y / originalHeight)
s1 = (start_x / originalWidth, start_y / originalHeight)
s2 = (start_x / originalWidth, end_y / originalHeight)
s3 = (end_x / originalWidth, start_y / originalHeight)
s4 = (end_x / originalWidth, end_y / originalHeight)
#print("points: ",p1,p2,p3,p4)
return p1,p2,p3,p4,start_x,start_y,end_x,end_y,cropWidth,cropHeight,originalSize
def writeCrop(id,coords,origin,crop,txt,start_x,start_y):
norm=[]
for point in coords:
normalized_x = ((point[0] * origin[0])-start_x)/crop[0]
normalized_y = ((point[1] * origin[1])-start_y)/ crop[1]
norm.append((normalized_x, normalized_y))
with open(txt, 'a') as file:
file.write(f"{id}")
for point in norm:
file.write(f" {point[0]} {point[1]}")
file.write("\n")
def crop(img_ratio,txt_path,image,output_path,scratch_ratio=0.8):
img = cv2.imread(image)
coords = []
with open(txt_path, 'r') as file:
for line in file:
point = []
parts = line.strip().split()
class_id = int(parts[0])
i = 1
while i + 1 < len(parts):
point.append((float(parts[i]), float(parts[i + 1])))
i += 2
coords.append(point)
scratch = Polygon(coords[0])
i = 0
while i < 5:
p1, p2, p3, p4, start_x, start_y, end_x, end_y, width, height, originalSize = pointGenCrop(img_ratio,img)
cropSize = [width, height]
points = [p1, p2, p3, p4]
promt = Polygon(points)
scratch_ar = scratch.area
if promt.intersection(scratch):
inter_ar = promt.intersection(scratch).area
if ((inter_ar / scratch_ar) > scratch_ratio):
cropped = img[start_y:end_y, start_x:end_x]
cv2.imwrite(output_path+ str(i) + ".jpg", cropped)
for temp in coords:
scratch = Polygon(temp)
inter = promt.intersection(scratch).exterior.coords
writeCrop(class_id, inter, originalSize, cropSize, output_path + str(i) + ".txt", start_x, start_y)
i += 1
#rotation
def rotate_image(img, angle,output_path):
# Rotate the image
rotated_img = img.rotate(angle, expand=False)
rotated_img.save(output_path+".jpg")
return rotated_img
def mult(angle, point,origin):
ox, oy = origin
px, py = point
qx = ox + math.cos(angle) * (px - ox) - math.sin(angle) * (py - oy)
qy = oy + math.sin(angle) * (px - ox) + math.cos(angle) * (py - oy)
return int(qx),int( qy)
def writeRot(coords, id, txt, width, height):
with open(txt, 'a') as file:
if len(coords)>1:
file.write(f"{id}")
for i, point in enumerate(coords):
if i < len(coords) - 1: # Exclude the last coordinate
file.write(f" {point[0] / width} {point[1] / height}")
file.write("\n")
def rotate(angle,txt_path,width,height,output_path):
rad=math.radians(angle)
rotatedTopLeft=mult(rad,(0,0),(width/2,height/2))
rotatedBottomRight=mult(rad,(width,height),(width/2,height/2))
rotatedFrame = [
(rotatedTopLeft[0], rotatedTopLeft[1]),
(rotatedTopLeft[0], rotatedBottomRight[1]),
(rotatedBottomRight[0], rotatedBottomRight[1]),
(rotatedBottomRight[0], rotatedTopLeft[1])
]
coords = []
rotatedCoords=[]
with open(txt_path, 'r') as file:
for line in file:
point = []
parts = line.strip().split()
class_id = int(parts[0])
i = 1
while i + 1 < len(parts):
point.append((float(parts[i]), float(parts[i + 1])))
i += 2
coords.append(point)
#print("point:",point)
for points in coords:
rotatedCoords=[]
for point in points:
rotatedCoords.append((mult(rad,(point[0]*width,point[1]*height),(width/2,height/2))))
framePoly = Polygon(rotatedFrame)
temp = Polygon(rotatedCoords).convex_hull
actual_frame = Polygon([(0, 0), (width, 0), (width, height), (0, height)])
intersection_polygon = actual_frame.intersection(temp)
finalScratch = intersection_polygon.exterior.coords
finalScratch_normalized = [(x / width, y / height) for x, y in finalScratch]
writeRot(finalScratch,class_id,output_path+".txt",width,height)
def executeRotation(image,angle,out,txt):
img = Image.open(image)
width, height = img.size
rotate_image(img,angle,out)
rotate(-angle,txt,width,height,out)
#color
#coordinate values between 0 and 1
def getNormCoord(txt_path):
coords = []
with open(txt_path, 'r') as file:
for line in file:
point = []
parts = line.strip().split()
class_id = int(parts[0])
i = 1
while i + 1 < len(parts):
point.append((float(parts[i]), float(parts[i + 1])))
i += 2
coords.append(point)
return coords
#pixel coordinates
def getPixelCor(coordinates,height,width):
got=[]
for (x,y) in coordinates:
got.append((int(x*width),int(y*height)))
return got
def find_most_common_color(image_path, polygon_vertices, color_intervals):
img = Image.open(image_path)
width, height = img.size
pixels = img.load()
polygon_path = Path(polygon_vertices)
color_counts = {color: 0 for color in color_intervals}
for y in range(height):
for x in range(width):
if polygon_path.contains_point((x, y)):
r, g, b = pixels[x, y]
for color, interval in color_intervals.items():
if interval[0] <= r <= interval[1] and \
interval[2] <= g <= interval[3] and \
interval[4] <= b <= interval[5]:
color_counts[color] += 1
print(color_counts)
most_common_color = max(color_counts, key=color_counts.get)
return most_common_color
def whiteChange(img, width, height, polygon_path, wg, pixels, color_intervals, output_path,image_name,mostcom):
i = 0
while i < wg:
m1 = random.random()
m2 = random.random()
m3 = random.random()
new_img = img.copy() # Create a copy of the original image
for y in range(height):
for x in range(width):
if polygon_path.contains_point((x, y)):
r, g, b = pixels[x, y]
for color, interval in color_intervals.items():
if interval[0] - 100 <= r <= interval[1] + 100 and \
interval[2] - 100 <= g <= interval[3] + 100 and \
interval[4] - 100 <= b <= interval[5] + 100:
if color == mostcom:
new_img.putpixel((x, y), (int(r * m1), int(g * m2), int(b * m3)))
new_img.save(os.path.join(output_path, image_name+str(i + 1)+'.png'))
i += 1
img.show()
def convert_to_color_orders(img, polygon_path,output_path,image_name):
img_array = np.array(img)
# Create a dictionary to map channel orders to their indices
channel_orders = {
'RGB': [0, 1, 2],
'RBG': [0, 2, 1],
'BGR': [2, 1, 0],
'BRG': [2, 0, 1],
'GBR': [1, 2, 0],
'GRB': [1, 0, 2]
}
i=0
# Iterate through each channel order and convert the image
for order_name, order_indices in channel_orders.items():
converted_img_array = img_array.copy()
for y in range(img.height):
for x in range(img.width):
if polygon_path.contains_point((x, y)):
r, g, b = img_array[y, x]
# Rearrange the values using order_indices
new_color = [0, 0, 0] # Initialize with zeros
new_color[order_indices[0]] = r
new_color[order_indices[1]] = g
new_color[order_indices[2]] = b
converted_img_array[y, x] = new_color
converted_img = Image.fromarray(converted_img_array)
converted_img.save(os.path.join(output_path, image_name+str(i + 1)+'.png'))
i+=1
def change_most_common_color(image_path, polygon_vertices, color_intervals,output_path,image_name,wg=1):
img = Image.open(image_path)
width, height = img.size
pixels = img.load()
polygon_path = Path(polygon_vertices)
most_common_color = find_most_common_color(image_path, polygon_vertices, color_intervals)
print(most_common_color)
if most_common_color == 'white':
whiteChange(img,width,height,polygon_path,wg,pixels,color_intervals,output_path,image_name,most_common_color)
elif most_common_color != 'black':
convert_to_color_orders(img,polygon_path,output_path,image_name)
#whiteChange(img,width,height,polygon_path,wg,pixels,color_intervals,output_path,image_name,most_common_color)
def lastcolor(model,image_path,output_path,wg=1):
color_intervals = {
'black': (0, 15, 0, 15, 0, 15),
'green': (0, 50, 50, 150, 0, 50),
'white': (175, 255, 175, 255, 175, 255),
'red': (10, 255, 0, 50, 0, 50),
'blue': (0, 50, 0, 80, 120, 255),
'orange': (150, 255, 50, 150, 0, 50)
}
full_filename = os.path.basename(image_path)
image_name = os.path.splitext(full_filename)[0]
image = cv2.imread(image_path)
height, width = image.shape[0], image.shape[1]
result = model(source=image, classes=[2], show_labels=False, show=True, line_width=1, save_txt=True,
save_conf=True,conf=0.5,iou=0.8)
save_path = result[-1].save_dir
txt_path = save_path + r"\labels\image0.txt"
a = getNormCoord(txt_path) # D:\PYTHON\demo1\Yolo task3\runs\segment\predict7\labels\image0.txt
t = getPixelCor(a[0], height, width)
change_most_common_color(image_path, t, color_intervals,output_path,image_name,wg)
###########EXAMPLE######################
#crop
#crop fonksiyonu ana fonksiyon
cropim=r"s.jpg"
croptxt=r"s.txt"
full_filename = os.path.basename(cropim)
img_name = os.path.splitext(full_filename)[0]
cropout = os.path.join(r"D:\PYTHON\demo1\Yolo task3\cropped", img_name + "_cropped")
img_ratio=0.5
crop_rat=0.8 #opsiyonel scratchin en az % kaçını alacağını gösterir
#crop(img_ratio,croptxt,cropim,cropout,crop_rat)
#rotation
rotim=r"D:\PYTHON\demo1\Yolo task3\torot\1164853167_right.jpeg"
rotxt=r"D:\PYTHON\demo1\Yolo task3\torot\1164853167_right.txt"
full_filename = os.path.basename(rotim)
image_name = os.path.splitext(full_filename)[0]
rotout = os.path.join(r"D:\PYTHON\demo1\Yolo task3\rotout", image_name + "_rotated")
ang=-50
#executeRotation(rotim,ang,rotout,rotxt)
#color
image_path = r"1165427507_left.jpeg"
out="D:\PYTHON\demo1\Yolo task3\colored"
#lastcolor(model,image_path,out,20)