-
Notifications
You must be signed in to change notification settings - Fork 1
/
app.py
379 lines (305 loc) · 14.7 KB
/
app.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
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import streamlit as st
import cv2
import numpy as np
from ultralytics import YOLO
from sort.sort import *
from tensorflow.keras.models import load_model
from util import get_car, read_license_plate, write_csv
import statistics as stats
# Set the GPU device
# import torch
# torch.cuda.set_device(0) # Set to your desired GPU number
# Load YOLO models
coco_model = YOLO('yolo_models/yolov8n.pt')
license_plate_detector = YOLO('yolo_models/license_plate_detector.pt')
# Load the character recognition model
model = load_model('updated_tfkmodel.keras')
# Initialize SORT tracker
mot_tracker = Sort()
def process_image(image):
# detect vehicles
detections = coco_model(image)[0]
detections_ = []
for detection in detections.boxes.data.tolist():
x1, y1, x2, y2, score, class_id = detection
# Considering only certain classes as vehicles
vehicles = [2, 3, 5, 7]
if int(class_id) in vehicles:
detections_.append([x1, y1, x2, y2, score])
# track vehicles
track_ids = mot_tracker.update(np.asarray(detections_))
# detect license plates
license_plates = license_plate_detector(image)[0]
license_plate_results = []
lp_crop = None
for license_plate in license_plates.boxes.data.tolist():
x1, y1, x2, y2, score, class_id = license_plate
# crop license plate from the original image
license_plate_crop_original = image[int(y1):int(y2), int(x1):int(x2), :]
lp_crop = license_plate_crop_original
# assign license plate to car
xcar1, ycar1, xcar2, ycar2, car_id = get_car(license_plate, track_ids)
if car_id != -1:
license_plate_results.append((car_id, license_plate_crop_original))
else:
continue
return license_plate_results, lp_crop
# Match contours to license plate or character template
def find_contours(dimensions, img):
# Find all contours in the image
cntrs, _ = cv2.findContours(img.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
# Retrieve potential dimensions
lower_width = dimensions[0]
upper_width = dimensions[1]
lower_height = dimensions[2]
upper_height = dimensions[3]
# Check largest 5 or 15 contours for license plate or character respectively
cntrs = sorted(cntrs, key=cv2.contourArea, reverse=True)[:15]
ii = cv2.imread('contour.jpg')
x_cntr_list = []
img_res = []
widths = []
heights = []
contours = [] # Store contour coordinates here
# Calculate the middle line of the license plate
middle_line = img.shape[0] // 2
# Sort contours based on x-coordinate and then on y-coordinate
sorted_cntrs = sorted(cntrs, key=lambda c: (c[0][0][0], c[0][0][1]))
# Separate contours above and below the middle line
above_middle = []
below_middle = []
for cntr in sorted_cntrs:
x, y, w, h = cv2.boundingRect(cntr)
if y < middle_line:
above_middle.append((x, y, w, h, cntr))
else:
below_middle.append((x, y, w, h, cntr))
# Sort contours from left to right
above_middle = sorted(above_middle, key=lambda c: c[0])
below_middle = sorted(below_middle, key=lambda c: c[0])
# Process contours above the middle line
for x, y, w, h, cntr in above_middle:
if w > lower_width and w < upper_width and h > lower_height and h < upper_height:
char = img[y:y+h, x:x+w]
white_pixels = np.sum(char == 255)
total_pixels = char.size
white_percentage = (white_pixels / total_pixels) * 100
if white_percentage >= 25:
x_cntr_list.append((x, y)) # stores the (x, y) coordinates of the character's contour
widths.append(w)
heights.append(h)
contours.append((x, y, x+w, y+h)) # Append contour coordinates
char_copy = np.zeros((44, 24))
# Extracting each character using the enclosing rectangle's coordinates.
char = cv2.resize(char, (20, 40))
cv2.rectangle(ii, (x, y), (x+w, y+h), (50, 21, 200), 2)
plt.imshow(ii, cmap='gray')
# Make result formatted for classification: invert colors
char = cv2.subtract(255, char)
# Resize the image to 24x44 with a black border
char_copy[2:42, 2:22] = char
char_copy[0:2, :] = 0
char_copy[:, 0:2] = 0
char_copy[42:44, :] = 0
char_copy[:, 22:24] = 0
img_res.append(char_copy) # List that stores the character's binary image (unsorted)
# Process contours below the middle line
for x, y, w, h, cntr in below_middle:
if w > lower_width and w < upper_width and h > lower_height and h < upper_height:
char = img[y:y+h, x:x+w]
white_pixels = np.sum(char == 255)
total_pixels = char.size
white_percentage = (white_pixels / total_pixels) * 100
if white_percentage >= 25:
x_cntr_list.append((x, y)) # stores the (x, y) coordinates of the character's contour
widths.append(w)
heights.append(h)
contours.append((x, y, x+w, y+h)) # Append contour coordinates
char_copy = np.zeros((44, 24))
# Extracting each character using the enclosing rectangle's coordinates.
char = cv2.resize(char, (20, 40))
cv2.rectangle(ii, (x, y), (x+w, y+h), (50, 21, 200), 2)
plt.imshow(ii, cmap='gray')
# Make result formatted for classification: invert colors
char = cv2.subtract(255, char)
# Resize the image to 24x44 with a black border
char_copy[2:42, 2:22] = char
char_copy[0:2, :] = 0
char_copy[:, 0:2] = 0
char_copy[42:44, :] = 0
char_copy[:, 22:24] = 0
img_res.append(char_copy) # List that stores the character's binary image (unsorted)
try:
# Calculate the mode of the widths and heights
mode_width = stats.mode(widths)
mode_height = stats.mode(heights)
print(f"Mode width: {mode_width}, Mode height: {mode_height}")
# Calculate median width and height
median_width = stats.median(widths)
median_height = stats.median(heights)
print(f"Median width: {median_width}, Median height: {median_height}")
# Calculate the mean of the widths and heights
mean_width = stats.mean(widths)
mean_height = stats.mean(heights)
# mean_width = (mode_width + median_width) / 2
# mean_height = (mode_height + median_height) / 2
print(f"Mean width: {mean_width}, Mean height: {mean_height}")
print(f"No# of characters: {len(img_res)}")
except:
print("The image quality is too low to detect characters.")
# Filter characters based on width and height deviation from the median
filtered_img_res = []
filtered_contours = []
for char, contour in zip(img_res, contours):
x1, y1, x2, y2 = contour
if ((x2 - x1) >= 0.70 * median_width) and ((x2 - x1) <= 1.3 * median_width) and ((y2 - y1) >= 0.70 * median_height) and ((y2 - y1) <= 1.3 * median_height):
filtered_img_res.append(char)
filtered_contours.append(contour)
# Remove contours with a distance of more than 15 pixels between them
remaining_contours = []
remaining_filtered_img_res = []
if len(filtered_contours) > 1:
center_x = img.shape[1] / 3
distance = 0
count = 1
for i in range(1, len(filtered_contours)):
if count >= len(filtered_contours):
break
x1_prev, _, x2_prev, _ = filtered_contours[count - 1]
x1_curr, _, x2_curr, _ = filtered_contours[count]
distance = x1_curr - x2_prev
print (f"Distance between contours {count} and {count+1}: {distance}")
if distance <= -175:
count += 1
continue
elif distance <= 15:
print("if (distance <= 15):")
remaining_contours.append(filtered_contours[count - 1])
remaining_filtered_img_res.append(filtered_img_res[count - 1])
count += 1
elif x1_curr < center_x:
print("elif x1_curr < center_x:")
remaining_contours.append(filtered_contours[count])
remaining_filtered_img_res.append(filtered_img_res[count])
count +=2
elif x1_curr >= (center_x * 2):
print("elif x1_curr >= center_x:")
remaining_contours.append(filtered_contours[count - 1])
remaining_filtered_img_res.append(filtered_img_res[count - 1])
count +=2
else:
print("else:")
remaining_contours.append(filtered_contours[count - 1])
remaining_filtered_img_res.append(filtered_img_res[count - 1])
count += 1
print("Last contour")
remaining_contours.append(filtered_contours[-1])
remaining_filtered_img_res.append(filtered_img_res[-1])
# plt.show()
return np.array(remaining_filtered_img_res)#, filtered_contours, remaining_contours
# Segment characters from the license plate
def segment_characters(image) :
# Preprocess cropped license plate image
img_lp = cv2.resize(image, (333, 75))
img_gray_lp = cv2.cvtColor(img_lp, cv2.COLOR_BGR2GRAY)
_, img_binary_lp = cv2.threshold(img_gray_lp, 200, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)
img_binary_lp = cv2.erode(img_binary_lp, (3,3))
img_binary_lp = cv2.dilate(img_binary_lp, (3,3))
LP_WIDTH = img_binary_lp.shape[0]
LP_HEIGHT = img_binary_lp.shape[1]
# Make borders white
img_binary_lp[0:3,:] = 255
img_binary_lp[:,0:3] = 255
img_binary_lp[72:75,:] = 255
img_binary_lp[:,330:333] = 255
# # Make borders white
# img_binary_lp[0:8,:] = 255
# img_binary_lp[:,0:25] = 255
# img_binary_lp[67:75,:] = 255
# img_binary_lp[:,308:333] = 255
# Estimations of character contours sizes of cropped license plates
dimensions = [LP_WIDTH/10,
2*LP_WIDTH/2.5,
LP_HEIGHT/20,
2*LP_HEIGHT/2.5]
# plt.show()
cv2.imwrite('contour.jpg', img_binary_lp)
# Get contours within cropped license plate
char_list = find_contours(dimensions, img_binary_lp)
# char_list, contours, remaining_contours = find_contours(dimensions, img_binary_lp)
return char_list#, contours, remaining_contours
def fix_dimension(img):
new_img = np.zeros((28,28,3))
for i in range(3):
new_img[:,:,i] = img
return new_img
def show_results(char_list, model):
dic = {}
characters = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'
for i,c in enumerate(characters):
dic[i] = c
output = []
for i,ch in enumerate(char_list): # iterating over the characters
img_ = cv2.resize(ch, (28,28))
img = fix_dimension(img_)
img = img.reshape(1,28,28,3) # preparing image for the model
# y_ = model.predict_classes(img)[0] # predicting the class
predict_x= model.predict(img)
classes_x = int(np.argmax(predict_x, axis=1)[0])
character = dic[classes_x] #
output.append(character) # storing the result in a list
plate_number = ''.join(output)
return plate_number
def main():
st.title("License Plate Detection and Character Recognition")
st.write('As a culminating project for the **Data Science and Machine Learning Bootcamp** at **Knowledge Streams**, we, [Umer Mansoor](https://www.linkedin.com/in/techsavvyumer/), [Wajahat Siddique](https://www.linkedin.com/in/wajahatsiddique/), and [Nayab Bashir](https://www.linkedin.com/in/nayabbashir/), under the mentorship of [Dr. Usman Nazir](https://www.linkedin.com/in/unazir/) and [Dr. Sana Jabbar](https://www.linkedin.com/in/sana-jabbar-9802a3252/), embarked on a journey to build a robust LPDR system.')
st.write('Through this project, we translated the theoretical and methodological knowledge gained at [Knowledge Streams](https://knowledge.tech/) into a tangible LPDR solution, solidifying our understanding and capabilities.')
uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
if uploaded_file is not None:
# Read the image
image = cv2.imdecode(np.frombuffer(uploaded_file.read(), np.uint8), 1)
# Process the image
lp_results, lp_img_crop = process_image(image)
# Display results
st.image(uploaded_file, caption="Uploaded Image", use_column_width=True)
if lp_results:
for _, license_plate_crop in lp_results:
st.image(license_plate_crop, caption=f"License Plate", use_column_width=True)
# Segment and recognize characters
char_list = segment_characters(license_plate_crop)
lp_number = show_results(char_list, model)
# for i in range(len(char_list)):
# plt.subplot(1, len(char_list), i+1)
# plt.imshow(char_list[i], cmap='gray')
# plt.axis('off')
# plt.savefig('segmented_chars.jpg')
if lp_number != "":
st.write(f"Recognized License Plate: {lp_number}")
else:
st.warning("The image quality is too low to detect characters.")
else:
if lp_img_crop is not None:
st.image(lp_img_crop, caption=f"License Plate", use_column_width=True)
char_list = segment_characters(lp_img_crop)
lp_number = show_results(char_list, model)
else:
st.image(image, caption=f"License Plate", use_column_width=True)
char_list = segment_characters(image)
lp_number = show_results(char_list, model)
# Segment and recognize characters
# char_list = segment_characters(lp_img_crop)
# lp_number = show_results(char_list, model)
# for i in range(len(char_list)):
# plt.subplot(1, len(char_list), i+1)
# plt.imshow(char_list[i], cmap='gray')
# plt.axis('off')
# plt.savefig('segmented_chars.jpg')
if lp_number != "":
st.write(f"Recognized License Plate: {lp_number}")
else:
st.warning("The image quality is too low to detect characters.")
if __name__ == "__main__":
main()