Repository: duanshengliu/End-to-end-for-chinese-plate-recognition
Branch: master
Commit: e47edd4aaf1f
Files: 12
Total size: 19.5 MB
Directory structure:
gitextract__ge8bvgl/
├── .gitignore
├── LICENSE
├── License-plate-recognition/
│ ├── CNN.py
│ ├── UI.py
│ ├── Unet.py
│ ├── __init__.py
│ ├── cnn.h5
│ ├── core.py
│ ├── train.py
│ └── unet.h5
├── README.md
└── requirements.txt
================================================
FILE CONTENTS
================================================
================================================
FILE: .gitignore
================================================
__pycache__/
================================================
FILE: LICENSE
================================================
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
Copyright 2023 duanshengliu
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
================================================
FILE: License-plate-recognition/CNN.py
================================================
# -*- coding:utf-8 -*-
# author: DuanshengLiu
from tensorflow.keras import layers, losses, models
import numpy as np
import cv2
import os
def cnn_train():
char_dict = {"京": 0, "沪": 1, "津": 2, "渝": 3, "冀": 4, "晋": 5, "蒙": 6, "辽": 7, "吉": 8, "黑": 9, "苏": 10,
"浙": 11, "皖": 12, "闽": 13, "赣": 14, "鲁": 15, "豫": 16, "鄂": 17, "湘": 18, "粤": 19, "桂": 20,
"琼": 21, "川": 22, "贵": 23, "云": 24, "藏": 25, "陕": 26, "甘": 27, "青": 28, "宁": 29, "新": 30,
"0": 31, "1": 32, "2": 33, "3": 34, "4": 35, "5": 36, "6": 37, "7": 38, "8": 39, "9": 40,
"A": 41, "B": 42, "C": 43, "D": 44, "E": 45, "F": 46, "G": 47, "H": 48, "J": 49, "K": 50,
"L": 51, "M": 52, "N": 53, "P": 54, "Q": 55, "R": 56, "S": 57, "T": 58, "U": 59, "V": 60,
"W": 61, "X": 62, "Y": 63, "Z": 64}
# 读取数据集
path = 'home/cnn_datasets/' # 车牌号数据集路径(车牌图片宽240，高80)
pic_name = sorted(os.listdir(path))
n = len(pic_name)
X_train, y_train = [], []
for i in range(n):
print("正在读取第%d张图片" % i)
img = cv2.imdecode(np.fromfile(path + pic_name[i], dtype=np.uint8), -1) # cv2.imshow无法读取中文路径图片，改用此方式
label = [char_dict[name] for name in pic_name[i][0:7]] # 图片名前7位为车牌标签
X_train.append(img)
y_train.append(label)
X_train = np.array(X_train)
y_train = [np.array(y_train)[:, i] for i in range(7)] # y_train是长度为7的列表，其中每个都是shape为(n,)的ndarray，分别对应n张图片的第一个字符，第二个字符....第七个字符
# cnn模型
Input = layers.Input((80, 240, 3)) # 车牌图片shape(80,240,3)
x = Input
x = layers.Conv2D(filters=16, kernel_size=(3, 3), strides=1, padding='same', activation='relu')(x)
x = layers.MaxPool2D(pool_size=(2, 2), padding='same', strides=2)(x)
for i in range(3):
x = layers.Conv2D(filters=32 * 2 ** i, kernel_size=(3, 3), padding='valid', activation='relu')(x)
x = layers.Conv2D(filters=32 * 2 ** i, kernel_size=(3, 3), padding='valid', activation='relu')(x)
x = layers.MaxPool2D(pool_size=(2, 2), padding='same', strides=2)(x)
x = layers.Dropout(0.5)(x)
x = layers.Flatten()(x)
x = layers.Dropout(0.3)(x)
Output = [layers.Dense(65, activation='softmax', name='c%d' % (i + 1))(x) for i in range(7)] # 7个输出分别对应车牌7个字符，每个输出都为65个类别类概率
model = models.Model(inputs=Input, outputs=Output)
model.summary()
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy', # y_train未进行one-hot编码，所以loss选择sparse_categorical_crossentropy
metrics=['accuracy'])
# 模型训练
print("开始训练cnn")
model.fit(X_train, y_train, epochs=35) # 总loss为7个loss的和
model.save('cnn.h5')
print('cnn.h5保存成功!!!')
def cnn_predict(cnn, Lic_img):
characters = ["京", "沪", "津", "渝", "冀", "晋", "蒙", "辽", "吉", "黑", "苏", "浙", "皖", "闽", "赣", "鲁", "豫",
"鄂", "湘", "粤", "桂", "琼", "川", "贵", "云", "藏", "陕", "甘", "青", "宁", "新", "0", "1", "2",
"3", "4", "5", "6", "7", "8", "9", "A", "B", "C", "D", "E", "F", "G", "H", "J", "K", "L", "M",
"N", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z"]
Lic_pred = []
for lic in Lic_img:
lic_pred = cnn.predict(lic.reshape(1, 80, 240, 3)) # 预测形状应为(1,80,240,3)
lic_pred = np.array(lic_pred).reshape(7, 65) # 列表转为ndarray，形状为(7,65)
if len(lic_pred[lic_pred >= 0.8]) >= 4: # 统计其中预测概率值大于80%以上的个数，大于等于4个以上认为识别率高，识别成功
chars = ''
for arg in np.argmax(lic_pred, axis=1): # 取每行中概率值最大的arg,将其转为字符
chars += characters[arg]
chars = chars[0:2] + '·' + chars[2:]
Lic_pred.append((lic, chars)) # 将车牌和识别结果一并存入Lic_pred
return Lic_pred
================================================
FILE: License-plate-recognition/UI.py
================================================
# -*- coding:utf-8 -*-
# author: DuanshengLiu
import cv2
import sys
import numpy as np
from tkinter import *
from tkinter.filedialog import askopenfilename
from PIL import Image, ImageTk
from tensorflow import keras
from core import locate_and_correct
from Unet import unet_predict
from CNN import cnn_predict
class Window:
def __init__(self, win, ww, wh):
self.win = win
self.ww = ww
self.wh = wh
self.win.geometry("%dx%d+%d+%d" % (ww, wh, 200, 50)) # 界面启动时的初始位置
self.win.title("车牌定位，矫正和识别软件---by DuanshengLiu")
self.img_src_path = None
self.label_src = Label(self.win, text='原图:', font=('微软雅黑', 13)).place(x=0, y=0)
self.label_lic1 = Label(self.win, text='车牌区域1:', font=('微软雅黑', 13)).place(x=615, y=0)
self.label_pred1 = Label(self.win, text='识别结果1:', font=('微软雅黑', 13)).place(x=615, y=85)
self.label_lic2 = Label(self.win, text='车牌区域2:', font=('微软雅黑', 13)).place(x=615, y=180)
self.label_pred2 = Label(self.win, text='识别结果2:', font=('微软雅黑', 13)).place(x=615, y=265)
self.label_lic3 = Label(self.win, text='车牌区域3:', font=('微软雅黑', 13)).place(x=615, y=360)
self.label_pred3 = Label(self.win, text='识别结果3:', font=('微软雅黑', 13)).place(x=615, y=445)
self.can_src = Canvas(self.win, width=512, height=512, bg='white', relief='solid', borderwidth=1) # 原图画布
self.can_src.place(x=50, y=0)
self.can_lic1 = Canvas(self.win, width=245, height=85, bg='white', relief='solid', borderwidth=1) # 车牌区域1画布
self.can_lic1.place(x=710, y=0)
self.can_pred1 = Canvas(self.win, width=245, height=65, bg='white', relief='solid', borderwidth=1) # 车牌识别1画布
self.can_pred1.place(x=710, y=90)
self.can_lic2 = Canvas(self.win, width=245, height=85, bg='white', relief='solid', borderwidth=1) # 车牌区域2画布
self.can_lic2.place(x=710, y=175)
self.can_pred2 = Canvas(self.win, width=245, height=65, bg='white', relief='solid', borderwidth=1) # 车牌识别2画布
self.can_pred2.place(x=710, y=265)
self.can_lic3 = Canvas(self.win, width=245, height=85, bg='white', relief='solid', borderwidth=1) # 车牌区域3画布
self.can_lic3.place(x=710, y=350)
self.can_pred3 = Canvas(self.win, width=245, height=65, bg='white', relief='solid', borderwidth=1) # 车牌识别3画布
self.can_pred3.place(x=710, y=440)
self.button1 = Button(self.win, text='选择文件', width=10, height=1, command=self.load_show_img) # 选择文件按钮
self.button1.place(x=680, y=wh - 30)
self.button2 = Button(self.win, text='识别车牌', width=10, height=1, command=self.display) # 识别车牌按钮
self.button2.place(x=780, y=wh - 30)
self.button3 = Button(self.win, text='清空所有', width=10, height=1, command=self.clear) # 清空所有按钮
self.button3.place(x=880, y=wh - 30)
self.unet = keras.models.load_model('unet.h5')
self.cnn = keras.models.load_model('cnn.h5')
print('正在启动中,请稍等...')
cnn_predict(self.cnn, [np.zeros((80, 240, 3))])
print("已启动,开始识别吧！")
def load_show_img(self):
self.clear()
sv = StringVar()
sv.set(askopenfilename())
self.img_src_path = Entry(self.win, state='readonly', text=sv).get() # 获取到所打开的图片
img_open = Image.open(self.img_src_path)
if img_open.size[0] * img_open.size[1] > 240 * 80:
img_open = img_open.resize((512, 512), Image.ANTIALIAS)
self.img_Tk = ImageTk.PhotoImage(img_open)
self.can_src.create_image(258, 258, image=self.img_Tk, anchor='center')
def display(self):
if self.img_src_path == None: # 还没选择图片就进行预测
self.can_pred1.create_text(32, 15, text='请选择图片', anchor='nw', font=('黑体', 28))
else:
img_src = cv2.imdecode(np.fromfile(self.img_src_path, dtype=np.uint8), -1) # 从中文路径读取时用
h, w = img_src.shape[0], img_src.shape[1]
if h * w <= 240 * 80 and 2 <= w / h <= 5: # 满足该条件说明可能整个图片就是一张车牌,无需定位,直接识别即可
lic = cv2.resize(img_src, dsize=(240, 80), interpolation=cv2.INTER_AREA)[:, :, :3] # 直接resize为(240,80)
img_src_copy, Lic_img = img_src, [lic]
else: # 否则就需通过unet对img_src原图预测,得到img_mask,实现车牌定位,然后进行识别
img_src, img_mask = unet_predict(self.unet, self.img_src_path)
img_src_copy, Lic_img = locate_and_correct(img_src, img_mask) # 利用core.py中的locate_and_correct函数进行车牌定位和矫正
Lic_pred = cnn_predict(self.cnn, Lic_img) # 利用cnn进行车牌的识别预测,Lic_pred中存的是元祖(车牌图片,识别结果)
if Lic_pred:
img = Image.fromarray(img_src_copy[:, :, ::-1]) # img_src_copy[:, :, ::-1]将BGR转为RGB
self.img_Tk = ImageTk.PhotoImage(img)
self.can_src.delete('all') # 显示前,先清空画板
self.can_src.create_image(258, 258, image=self.img_Tk,
anchor='center') # img_src_copy上绘制出了定位的车牌轮廓,将其显示在画板上
for i, lic_pred in enumerate(Lic_pred):
if i == 0:
self.lic_Tk1 = ImageTk.PhotoImage(Image.fromarray(lic_pred[0][:, :, ::-1]))
self.can_lic1.create_image(5, 5, image=self.lic_Tk1, anchor='nw')
self.can_pred1.create_text(35, 15, text=lic_pred[1], anchor='nw', font=('黑体', 28))
elif i == 1:
self.lic_Tk2 = ImageTk.PhotoImage(Image.fromarray(lic_pred[0][:, :, ::-1]))
self.can_lic2.create_image(5, 5, image=self.lic_Tk2, anchor='nw')
self.can_pred2.create_text(40, 15, text=lic_pred[1], anchor='nw', font=('黑体', 28))
elif i == 2:
self.lic_Tk3 = ImageTk.PhotoImage(Image.fromarray(lic_pred[0][:, :, ::-1]))
self.can_lic3.create_image(5, 5, image=self.lic_Tk3, anchor='nw')
self.can_pred3.create_text(40, 15, text=lic_pred[1], anchor='nw', font=('黑体', 28))
else: # Lic_pred为空说明未能识别
self.can_pred1.create_text(47, 15, text='未能识别', anchor='nw', font=('黑体', 27))
def clear(self):
self.can_src.delete('all')
self.can_lic1.delete('all')
self.can_lic2.delete('all')
self.can_lic3.delete('all')
self.can_pred1.delete('all')
self.can_pred2.delete('all')
self.can_pred3.delete('all')
self.img_src_path = None
def closeEvent(): # 关闭前清除session(),防止'NoneType' object is not callable
keras.backend.clear_session()
sys.exit()
if __name__ == '__main__':
win = Tk()
ww = 1000 # 窗口宽设定1000
wh = 600 # 窗口高设定600
Window(win, ww, wh)
win.protocol("WM_DELETE_WINDOW", Window.closeEvent)
win.mainloop()
================================================
FILE: License-plate-recognition/Unet.py
================================================
# -*- coding:utf-8 -*-
# author: DuanshengLiu
import numpy as np
import os
import cv2
from tensorflow.keras import layers, losses, models
def unet_train():
height = 512
width = 512
path = 'D:/desktop/unet_datasets/'
input_name = os.listdir(path + 'train_image')
n = len(input_name)
print(n)
X_train, y_train = [], []
for i in range(n):
print("正在读取第%d张图片" % i)
img = cv2.imread(path + 'train_image/%d.png' % i)
label = cv2.imread(path + 'train_label/%d.png' % i)
X_train.append(img)
y_train.append(label)
X_train = np.array(X_train)
y_train = np.array(y_train)
def Conv2d_BN(x, nb_filter, kernel_size, strides=(1, 1), padding='same'):
x = layers.Conv2D(nb_filter, kernel_size, strides=strides, padding=padding)(x)
x = layers.BatchNormalization(axis=3)(x)
x = layers.LeakyReLU(alpha=0.1)(x)
return x
def Conv2dT_BN(x, filters, kernel_size, strides=(2, 2), padding='same'):
x = layers.Conv2DTranspose(filters, kernel_size, strides=strides, padding=padding)(x)
x = layers.BatchNormalization(axis=3)(x)
x = layers.LeakyReLU(alpha=0.1)(x)
return x
inpt = layers.Input(shape=(height, width, 3))
conv1 = Conv2d_BN(inpt, 8, (3, 3))
conv1 = Conv2d_BN(conv1, 8, (3, 3))
pool1 = layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same')(conv1)
conv2 = Conv2d_BN(pool1, 16, (3, 3))
conv2 = Conv2d_BN(conv2, 16, (3, 3))
pool2 = layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same')(conv2)
conv3 = Conv2d_BN(pool2, 32, (3, 3))
conv3 = Conv2d_BN(conv3, 32, (3, 3))
pool3 = layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same')(conv3)
conv4 = Conv2d_BN(pool3, 64, (3, 3))
conv4 = Conv2d_BN(conv4, 64, (3, 3))
pool4 = layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same')(conv4)
conv5 = Conv2d_BN(pool4, 128, (3, 3))
conv5 = layers.Dropout(0.5)(conv5)
conv5 = Conv2d_BN(conv5, 128, (3, 3))
conv5 = layers.Dropout(0.5)(conv5)
convt1 = Conv2dT_BN(conv5, 64, (3, 3))
concat1 = layers.concatenate([conv4, convt1], axis=3)
concat1 = layers.Dropout(0.5)(concat1)
conv6 = Conv2d_BN(concat1, 64, (3, 3))
conv6 = Conv2d_BN(conv6, 64, (3, 3))
convt2 = Conv2dT_BN(conv6, 32, (3, 3))
concat2 = layers.concatenate([conv3, convt2], axis=3)
concat2 = layers.Dropout(0.5)(concat2)
conv7 = Conv2d_BN(concat2, 32, (3, 3))
conv7 = Conv2d_BN(conv7, 32, (3, 3))
convt3 = Conv2dT_BN(conv7, 16, (3, 3))
concat3 = layers.concatenate([conv2, convt3], axis=3)
concat3 = layers.Dropout(0.5)(concat3)
conv8 = Conv2d_BN(concat3, 16, (3, 3))
conv8 = Conv2d_BN(conv8, 16, (3, 3))
convt4 = Conv2dT_BN(conv8, 8, (3, 3))
concat4 = layers.concatenate([conv1, convt4], axis=3)
concat4 = layers.Dropout(0.5)(concat4)
conv9 = Conv2d_BN(concat4, 8, (3, 3))
conv9 = Conv2d_BN(conv9, 8, (3, 3))
conv9 = layers.Dropout(0.5)(conv9)
outpt = layers.Conv2D(filters=3, kernel_size=(1, 1), strides=(1, 1), padding='same', activation='relu')(conv9) # TODO:activation='softmax'
model = models.Model(inpt, outpt)
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
model.summary()
print("开始训练u-net")
model.fit(X_train, y_train, epochs=100, batch_size=15)#epochs和batch_size看个人情况调整，batch_size不要过大，否则内存容易溢出
#我11G显存也只能设置15-20左右，我训练最终loss降低至250左右，acc约95%左右
model.save('unet.h5')
print('unet.h5保存成功!!!')
def unet_predict(unet, img_src_path):
img_src = cv2.imdecode(np.fromfile(img_src_path, dtype=np.uint8), -1) # 从中文路径读取时用
# img_src=cv2.imread(img_src_path)
if img_src.shape != (512, 512, 3):
img_src = cv2.resize(img_src, dsize=(512, 512), interpolation=cv2.INTER_AREA)[:, :, :3] # dsize=(宽度,高度),[:,:,:3]是防止图片为4通道图片，后续无法reshape
img_src = img_src.reshape(1, 512, 512, 3) # 预测图片shape为(1,512,512,3)
img_mask = unet.predict(img_src) # 归一化除以255后进行预测
img_src = img_src.reshape(512, 512, 3) # 将原图reshape为3维
img_mask = img_mask.reshape(512, 512, 3) # 将预测后图片reshape为3维
img_mask = img_mask / np.max(img_mask) * 255 # 归一化后乘以255
img_mask[:, :, 2] = img_mask[:, :, 1] = img_mask[:, :, 0] # 三个通道保持相同
img_mask = img_mask.astype(np.uint8) # 将img_mask类型转为int型
return img_src, img_mask
================================================
FILE: License-plate-recognition/__init__.py
================================================
================================================
FILE: License-plate-recognition/cnn.h5
================================================
[File too large to display: 19.5 MB]
================================================
FILE: License-plate-recognition/core.py
================================================
# -*- coding:utf-8 -*-
# author: DuanshengLiu
import cv2
import numpy as np
def locate_and_correct(img_src, img_mask):
"""
该函数通过cv2对img_mask进行边缘检测，获取车牌区域的边缘坐标(存储在contours中)和最小外接矩形4个端点坐标,
再从车牌的边缘坐标中计算出和最小外接矩形4个端点最近的点即为平行四边形车牌的四个端点,从而实现车牌的定位和矫正
:param img_src: 原始图片
:param img_mask: 通过u_net进行图像分隔得到的二值化图片，车牌区域呈现白色，背景区域为黑色
:return: 定位且矫正后的车牌
"""
# cv2.imshow('img_mask',img_mask)
# cv2.waitKey(0)
# ret,thresh = cv2.threshold(img_mask[:,:,0],0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU) #二值化
# cv2.imshow('thresh',thresh)
# cv2.waitKey(0)
try:
contours, hierarchy = cv2.findContours(img_mask[:, :, 0], cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
except: # 防止opencv版本不一致报错
ret, contours, hierarchy = cv2.findContours(img_mask[:, :, 0], cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if not len(contours): # contours1长度为0说明未检测到车牌
# print("未检测到车牌")
return [], []
else:
Lic_img = []
img_src_copy = img_src.copy() # img_src_copy用于绘制出定位的车牌轮廓
for ii, cont in enumerate(contours):
x, y, w, h = cv2.boundingRect(cont) # 获取最小外接矩形
img_cut_mask = img_mask[y:y + h, x:x + w] # 将标签车牌区域截取出来
# cv2.imshow('img_cut_mask',img_cut_mask)
# cv2.waitKey(0)
# print('w,h,均值,宽高比',w,h,np.mean(img_cut_mask),w/h)
# contours中除了车牌区域可能会有宽或高都是1或者2这样的小噪点，
# 而待选车牌区域的均值应较高，且宽和高不会非常小，因此通过以下条件进行筛选
if np.mean(img_cut_mask) >= 75 and w > 15 and h > 15:
rect = cv2.minAreaRect(cont) # 针对坐标点获取带方向角的最小外接矩形，中心点坐标，宽高，旋转角度
box = cv2.boxPoints(rect).astype(np.int32) # 获取最小外接矩形四个顶点坐标
# cv2.drawContours(img_mask, contours, -1, (0, 0, 255), 2)
# cv2.drawContours(img_mask, [box], 0, (0, 255, 0), 2)
# cv2.imshow('img_mask',img_mask)
# cv2.waitKey(0)
cont = cont.reshape(-1, 2).tolist()
# 由于转换矩阵的两组坐标位置需要一一对应，因此需要将最小外接矩形的坐标进行排序，最终排序为[左上，左下，右上，右下]
box = sorted(box, key=lambda xy: xy[0]) # 先按照左右进行排序，分为左侧的坐标和右侧的坐标
box_left, box_right = box[:2], box[2:] # 此时box的前2个是左侧的坐标，后2个是右侧的坐标
box_left = sorted(box_left, key=lambda x: x[1]) # 再按照上下即y进行排序，此时box_left中为左上和左下两个端点坐标
box_right = sorted(box_right, key=lambda x: x[1]) # 此时box_right中为右上和右下两个端点坐标
box = np.array(box_left + box_right) # [左上，左下，右上，右下]
# print(box)
x0, y0 = box[0][0], box[0][1] # 这里的4个坐标即为最小外接矩形的四个坐标，接下来需获取平行(或不规则)四边形的坐标
x1, y1 = box[1][0], box[1][1]
x2, y2 = box[2][0], box[2][1]
x3, y3 = box[3][0], box[3][1]
def point_to_line_distance(X, Y):
if x2 - x0:
k_up = (y2 - y0) / (x2 - x0) # 斜率不为无穷大
d_up = abs(k_up * X - Y + y2 - k_up * x2) / (k_up ** 2 + 1) ** 0.5
else: # 斜率无穷大
d_up = abs(X - x2)
if x1 - x3:
k_down = (y1 - y3) / (x1 - x3) # 斜率不为无穷大
d_down = abs(k_down * X - Y + y1 - k_down * x1) / (k_down ** 2 + 1) ** 0.5
else: # 斜率无穷大
d_down = abs(X - x1)
return d_up, d_down
d0, d1, d2, d3 = np.inf, np.inf, np.inf, np.inf
l0, l1, l2, l3 = (x0, y0), (x1, y1), (x2, y2), (x3, y3)
for each in cont: # 计算cont中的坐标与矩形四个坐标的距离以及到上下两条直线的距离，对距离和进行权重的添加，成功计算选出四边形的4个顶点坐标
x, y = each[0], each[1]
dis0 = (x - x0) ** 2 + (y - y0) ** 2
dis1 = (x - x1) ** 2 + (y - y1) ** 2
dis2 = (x - x2) ** 2 + (y - y2) ** 2
dis3 = (x - x3) ** 2 + (y - y3) ** 2
d_up, d_down = point_to_line_distance(x, y)
weight = 0.975
if weight * d_up + (1 - weight) * dis0 < d0: # 小于则更新
d0 = weight * d_up + (1 - weight) * dis0
l0 = (x, y)
if weight * d_down + (1 - weight) * dis1 < d1:
d1 = weight * d_down + (1 - weight) * dis1
l1 = (x, y)
if weight * d_up + (1 - weight) * dis2 < d2:
d2 = weight * d_up + (1 - weight) * dis2
l2 = (x, y)
if weight * d_down + (1 - weight) * dis3 < d3:
d3 = weight * d_down + (1 - weight) * dis3
l3 = (x, y)
# print([l0,l1,l2,l3])
# for l in [l0, l1, l2, l3]:
# cv2.circle(img=img_mask, color=(0, 255, 255), center=tuple(l), thickness=2, radius=2)
# cv2.imshow('img_mask',img_mask)
# cv2.waitKey(0)
p0 = np.float32([l0, l1, l2, l3]) # 左上角，左下角，右上角，右下角，p0和p1中的坐标顺序对应，以进行转换矩阵的形成
p1 = np.float32([(0, 0), (0, 80), (240, 0), (240, 80)]) # 我们所需的长方形
transform_mat = cv2.getPerspectiveTransform(p0, p1) # 构成转换矩阵
lic = cv2.warpPerspective(img_src, transform_mat, (240, 80)) # 进行车牌矫正
# cv2.imshow('lic',lic)
# cv2.waitKey(0)
Lic_img.append(lic)
cv2.drawContours(img_src_copy, [np.array([l0, l1, l3, l2])], -1, (0, 255, 0), 2) # 在img_src_copy上绘制出定位的车牌轮廓，(0, 255, 0)表示绘制线条为绿色
return img_src_copy, Lic_img
================================================
FILE: License-plate-recognition/train.py
================================================
# -*- coding:utf-8 -*-
# author: DuanshengLiu
from Unet import unet_train
from CNN import cnn_train
unet_train()#训练后得到unet.h5,用于车牌定位
cnn_train()#训练后得到cnn.h5，用于车牌识别
================================================
FILE: README.md
================================================
# End-to-end-for-chinese-plate-recognition
## 基于u-net，cv2以及cnn的中文车牌定位，矫正和端到端识别软件，其中unet和cv2用于车牌定位和矫正，cnn进行车牌识别，unet和cnn都是基于tensorflow的keras实现
## 环境：python:3.6, tensorflow:1.15.2, opencv: 4.1.0.25, keras: 2.3.1
### 整体思路：1. 利用u-net图像分割得到二值化图像，2. 再使用cv2进行边缘检测获得车牌区域坐标，并将车牌图形矫正，3. 利用卷积神经网络cnn进行车牌多标签端到端识别，具体描述可见CSDN博客：https://blog.csdn.net/qq_32194791/article/details/106748685
### 实现效果：拍摄角度倾斜、强曝光或昏暗环境等都能较好地识别，甚至有些百度AI车牌识别未能识别的图片也能识别
### 注意：若是直接识别类似下图的无需定位的完整车牌，那么请确保图片尺寸小于等于240 * 80，否则会被认为图片中含其余区域而进行定位，反而识别效果不佳
### 其余的没什么问题，正常识别都可以
### 部分效果图：
================================================
FILE: requirements.txt
================================================
tensorflow==1.15.2
opencv-python==4.1.0.25
keras==2.3.1
pillow==9.0.0
gitextract__ge8bvgl/ ├── .gitignore ├── LICENSE ├── License-plate-recognition/ │ ├── CNN.py │ ├── UI.py │ ├── Unet.py │ ├── __init__.py │ ├── cnn.h5 │ ├── core.py │ ├── train.py │ └── unet.h5 ├── README.md └── requirements.txt
SYMBOL INDEX (11 symbols across 4 files)
FILE: License-plate-recognition/CNN.py
function cnn_train (line 9) | def cnn_train():
function cnn_predict (line 58) | def cnn_predict(cnn, Lic_img):
FILE: License-plate-recognition/UI.py
class Window (line 15) | class Window:
method __init__ (line 16) | def __init__(self, win, ww, wh):
method load_show_img (line 60) | def load_show_img(self):
method display (line 71) | def display(self):
method clear (line 108) | def clear(self):
method closeEvent (line 118) | def closeEvent(): # 关闭前清除session(),防止'NoneType' object is not callable
FILE: License-plate-recognition/Unet.py
function unet_train (line 9) | def unet_train():
function unet_predict (line 100) | def unet_predict(unet, img_src_path):
FILE: License-plate-recognition/core.py
function locate_and_correct (line 7) | def locate_and_correct(img_src, img_mask):
Condensed preview — 12 files, each showing path, character count, and a content snippet. Download the .json file or copy for the full structured content (39K chars).
[
{
"path": ".gitignore",
"chars": 12,
"preview": "__pycache__/"
},
{
"path": "LICENSE",
"chars": 10756,
"preview": " Apache License\n Version 2.0, January 2004\n "
},
{
"path": "License-plate-recognition/CNN.py",
"chars": 3768,
"preview": "# -*- coding:utf-8 -*-\r\n# author: DuanshengLiu\r\nfrom tensorflow.keras import layers, losses, models\r\nimport numpy as np\r"
},
{
"path": "License-plate-recognition/UI.py",
"chars": 6869,
"preview": "# -*- coding:utf-8 -*-\r\n# author: DuanshengLiu\r\nimport cv2\r\nimport sys\r\nimport numpy as np\r\nfrom tkinter import *\r\nfrom "
},
{
"path": "License-plate-recognition/Unet.py",
"chars": 4572,
"preview": "# -*- coding:utf-8 -*-\r\n# author: DuanshengLiu\r\nimport numpy as np\r\nimport os\r\nimport cv2\r\nfrom tensorflow.keras import "
},
{
"path": "License-plate-recognition/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "License-plate-recognition/core.py",
"chars": 5673,
"preview": "# -*- coding:utf-8 -*-\r\n# author: DuanshengLiu\r\nimport cv2\r\nimport numpy as np\r\n\r\n\r\ndef locate_and_correct(img_src, img_"
},
{
"path": "License-plate-recognition/train.py",
"chars": 172,
"preview": "# -*- coding:utf-8 -*-\r\n# author: DuanshengLiu\r\nfrom Unet import unet_train\r\nfrom CNN import cnn_train\r\n\r\nunet_train()#训"
},
{
"path": "README.md",
"chars": 1491,
"preview": "# End-to-end-for-chinese-plate-recognition\n\n## 基于u-net，cv2以及cnn的中文车牌定位，矫正和端到端识别软件，其中unet和cv2用于车牌定位和矫正，cnn进行车牌识别，unet和cnn"
},
{
"path": "requirements.txt",
"chars": 69,
"preview": "tensorflow==1.15.2\nopencv-python==4.1.0.25\nkeras==2.3.1\npillow==9.0.0"
}
]
// ... and 2 more files (download for full content)
About this extraction
This page contains the full source code of the duanshengliu/End-to-end-for-chinese-plate-recognition GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 12 files (19.5 MB), approximately 10.7k tokens, and a symbol index with 11 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.
Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.