【Python实战】——Python+Opencv是实现车牌自动识别
小墨&晓末 2024-07-29 08:05:11 阅读 56
🍉CSDN小墨&晓末:https://blog.csdn.net/jd1813346972
个人介绍: 研一|统计学|干货分享
擅长Python、Matlab、R等主流编程软件
累计十余项国家级比赛奖项,参与研究经费10w、40w级横向
文章目录
1 导入相关模块2 相关功能函数定义2.1 彩色图片显示函数(plt_show0)2.2 灰度图片显示函数(plt_show)2.3 图像去噪函数(gray_guss)
2 图像预处理2.1 图片读取2.2 高斯去噪2.3 边缘检测
2.4 阈值化3 车牌定位3.1 区域选择3.2 形态学操作3.3 轮廓检测
4 车牌字符分割4.1 高斯去噪4.2 阈值化4.3 膨胀操作4.4 车牌号排序4.5 分割效果
5 模板匹配5.1 准备模板5.2 匹配结果5.3 匹配效果展示
6完整代码
该篇文章将以实战形式演示利用Python结合Opencv实现车牌识别,全程涉及图像预处理、车牌定位、车牌分割、通过模板匹配识别结果输出。该项目对于智能交通、车辆管理等领域具有实际应用价值。通过自动识别车牌号码,可以实现车辆追踪、违章查询、停车场管理等功能,提高交通管理的效率和准确性。可用于车牌识别技术学习。
技术要点:
OpenCV:用于图像处理和计算机视觉任务。Python:作为编程语言,具有简单易学、资源丰富等优点。图像处理技术:如灰度化、噪声去除、边缘检测、形态学操作、透视变换等。
1 导入相关模块
<code>import cv2
from matplotlib import pyplot as plt
import os
import numpy as np
from PIL import ImageFont, ImageDraw, Image
2 相关功能函数定义
2.1 彩色图片显示函数(plt_show0)
def plt_show0(img):
b,g,r = cv2.split(img)
img = cv2.merge([r, g, b])
plt.imshow(img)
plt.show()
cv2与plt的图像通道不同:cv2为[b,g,r];plt为[r, g, b]
2.2 灰度图片显示函数(plt_show)
def plt_show(img):
plt.imshow(img,cmap='gray')code>
plt.show()
2.3 图像去噪函数(gray_guss)
def gray_guss(image):
image = cv2.GaussianBlur(image, (3, 3), 0)
gray_image = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
return gray_image
此处演示使用高斯模糊去噪。
cv2.GaussianBlur参数说明:
src
:输入图像,可以是任意数量的通道,这些通道可以独立处理,但深度应为 CV_8U
、CV_16U
、CV_16S
、CV_32F
或 CV_64F
。ksize
:高斯核的大小,必须是正奇数,例如 (3, 3)、(5, 5) 等。如果 ksize
的值为零,那么它会根据 sigmaX
和 sigmaY
的值来计算。sigmaX
:X 方向上的高斯核标准偏差。dst
:输出图像,大小和类型与 src
相同。sigmaY
:Y 方向上的高斯核标准偏差,如果 sigmaY
是零,那么它会与 sigmaX
的值相同。如果 sigmaY
是负数,那么它会从 ksize.width
和 ksize.height
计算得出。borderType
:像素外插法,有默认值。
2 图像预处理
2.1 图片读取
origin_image = cv2.imread('D:/image/car3.jpg')
此处演示识别车牌原图:
2.2 高斯去噪
<code>origin_image = cv2.imread('D:/image/car3.jpg')
# 复制一张图片,在复制图上进行图像操作,保留原图
image = origin_image.copy()
gray_image = gray_guss(image)
2.3 边缘检测
Sobel_x = cv2.Sobel(gray_image, cv2.CV_16S, 1, 0)
absX = cv2.convertScaleAbs(Sobel_x)
image = absX
x方向上的边缘检测(增强边缘信息)。
2.4 阈值化
# 图像阈值化操作——获得二值化图
ret, image = cv2.threshold(image, 0, 255, cv2.THRESH_OTSU)
# 显示灰度图像
plt_show(image)
运行结果:
3 车牌定位
3.1 区域选择
<code>kernelX = cv2.getStructuringElement(cv2.MORPH_RECT, (30, 10))
image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernelX,iterations = 1)
# 显示灰度图像
plt_show(image)
从图像中提取对表达和描绘区域形状有意义的图像分量。
运行结果:
3.2 形态学操作
<code># 腐蚀(erode)和膨胀(dilate)
kernelX = cv2.getStructuringElement(cv2.MORPH_RECT, (50, 1))
kernelY = cv2.getStructuringElement(cv2.MORPH_RECT, (1, 20))
#x方向进行闭操作(抑制暗细节)
image = cv2.dilate(image, kernelX)
image = cv2.erode(image, kernelX)
#y方向的开操作
image = cv2.erode(image, kernelY)
image = cv2.dilate(image, kernelY)
# 中值滤波(去噪)
image = cv2.medianBlur(image, 21)
# 显示灰度图像
plt_show(image)
使用膨胀和腐蚀操作来突出车牌区域。
运行结果:
3.3 轮廓检测
<code>contours, hierarchy = cv2.findContours(image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
for item in contours:
rect = cv2.boundingRect(item)
x = rect[0]
y = rect[1]
weight = rect[2]
height = rect[3]
# 根据轮廓的形状特点,确定车牌的轮廓位置并截取图像
if (weight > (height * 3)) and (weight < (height * 4.5)):
image = origin_image[y:y + height, x:x + weight]
plt_show(image)
4 车牌字符分割
4.1 高斯去噪
<code># 图像去噪灰度处理
gray_image = gray_guss(image)
4.2 阈值化
ret, image = cv2.threshold(gray_image, 0, 255, cv2.THRESH_OTSU)
plt_show(image)
运行结果:
4.3 膨胀操作
<code>#膨胀操作
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (4, 4))
image = cv2.dilate(image, kernel)
plt_show(image)
运行结果:
4.4 车牌号排序
<code>words = sorted(words,key=lambda s:s[0],reverse=False)
i = 0
#word中存放轮廓的起始点和宽高
for word in words:
# 筛选字符的轮廓
if (word[3] > (word[2] * 1.5)) and (word[3] < (word[2] * 5.5)) and (word[2] > 10):
i = i+1
if word[2] < 15:
splite_image = image[word[1]:word[1] + word[3], word[0]-word[2]:word[0] + word[2]*2]
else:
splite_image = image[word[1]:word[1] + word[3], word[0]:word[0] + word[2]]
word_images.append(splite_image)
print(i)
print(words)
运行结果:
1
2
3
4
5
6
7
[[2, 0, 7, 70], [12, 6, 30, 55], [15, 7, 7, 9], [46, 6, 32, 55], [83, 30, 9, 9], [96, 7, 32, 55], [132, 8, 32, 55], [167, 8, 30, 54], [202, 62, 7, 6], [203, 7, 30, 55], [245, 7, 12, 54], [266, 0, 12, 70]]
4.5 分割效果
for i,j in enumerate(word_images):
plt.subplot(1,7,i+1)
plt.imshow(word_images[i],cmap='gray')code>
plt.show()
运行结果:
5 模板匹配
5.1 准备模板
<code># 准备模板(template[0-9]为数字模板;)
template = ['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',
'藏','川','鄂','甘','赣','贵','桂','黑','沪','吉','冀','津','晋','京','辽','鲁','蒙','闽','宁',
'青','琼','陕','苏','皖','湘','新','渝','豫','粤','云','浙']
# 读取一个文件夹下的所有图片,输入参数是文件名,返回模板文件地址列表
def read_directory(directory_name):
referImg_list = []
for filename in os.listdir(directory_name):
referImg_list.append(directory_name + "/" + filename)
return referImg_list
# 获得中文模板列表(只匹配车牌的第一个字符)
def get_chinese_words_list():
chinese_words_list = []
for i in range(34,64):
#将模板存放在字典中
c_word = read_directory('D:/refer1/'+ template[i])
chinese_words_list.append(c_word)
return chinese_words_list
chinese_words_list = get_chinese_words_list()
# 获得英文模板列表(只匹配车牌的第二个字符)
def get_eng_words_list():
eng_words_list = []
for i in range(10,34):
e_word = read_directory('D:/refer1/'+ template[i])
eng_words_list.append(e_word)
return eng_words_list
eng_words_list = get_eng_words_list()
# 获得英文和数字模板列表(匹配车牌后面的字符)
def get_eng_num_words_list():
eng_num_words_list = []
for i in range(0,34):
word = read_directory('D:/refer1/'+ template[i])
eng_num_words_list.append(word)
return eng_num_words_list
eng_num_words_list = get_eng_num_words_list()
此处需提前准备各类字符模板。
5.2 匹配结果
# 获得英文和数字模板列表(匹配车牌后面的字符)
def get_eng_num_words_list():
eng_num_words_list = []
for i in range(0,34):
word = read_directory('D:/refer1/'+ template[i])
eng_num_words_list.append(word)
return eng_num_words_list
eng_num_words_list = get_eng_num_words_list()
# 读取一个模板地址与图片进行匹配,返回得分
def template_score(template,image):
#将模板进行格式转换
template_img=cv2.imdecode(np.fromfile(template,dtype=np.uint8),1)
template_img = cv2.cvtColor(template_img, cv2.COLOR_RGB2GRAY)
#模板图像阈值化处理——获得黑白图
ret, template_img = cv2.threshold(template_img, 0, 255, cv2.THRESH_OTSU)
# height, width = template_img.shape
# image_ = image.copy()
# image_ = cv2.resize(image_, (width, height))
image_ = image.copy()
#获得待检测图片的尺寸
height, width = image_.shape
# 将模板resize至与图像一样大小
template_img = cv2.resize(template_img, (width, height))
# 模板匹配,返回匹配得分
result = cv2.matchTemplate(image_, template_img, cv2.TM_CCOEFF)
return result[0][0]
# 对分割得到的字符逐一匹配
def template_matching(word_images):
results = []
for index,word_image in enumerate(word_images):
if index==0:
best_score = []
for chinese_words in chinese_words_list:
score = []
for chinese_word in chinese_words:
result = template_score(chinese_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[34+i])
r = template[34+i]
results.append(r)
continue
if index==1:
best_score = []
for eng_word_list in eng_words_list:
score = []
for eng_word in eng_word_list:
result = template_score(eng_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[10+i])
r = template[10+i]
results.append(r)
continue
else:
best_score = []
for eng_num_word_list in eng_num_words_list:
score = []
for eng_num_word in eng_num_word_list:
result = template_score(eng_num_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[i])
r = template[i]
results.append(r)
continue
return results
word_images_ = word_images.copy()
# 调用函数获得结果
result = template_matching(word_images_)
print(result)
print( "".join(result))
运行结果:
['渝', 'B', 'F', 'U', '8', '7', '1']
渝BFU871
“”.join(result)函数将列表转换为拼接好的字符串,方便结果显示
5.3 匹配效果展示
height,weight = origin_image.shape[0:2]
print(height)
print(weight)
image_1 = origin_image.copy()
cv2.rectangle(image_1, (int(0.2*weight), int(0.75*height)), (int(weight*0.9), int(height*0.95)), (0, 255, 0), 5)
#设置需要显示的字体
fontpath = "font/simsun.ttc"
font = ImageFont.truetype(fontpath,64)
img_pil = Image.fromarray(image_1)
draw = ImageDraw.Draw(img_pil)
#绘制文字信息
draw.text((int(0.2*weight)+25, int(0.75*height)), "".join(result), font = font, fill = (255, 255, 0))
bk_img = np.array(img_pil)
print(result)
print( "".join(result))
plt_show0(bk_img)
运行结果:
6完整代码
<code># 导入所需模块
import cv2
from matplotlib import pyplot as plt
import os
import numpy as np
from PIL import ImageFont, ImageDraw, Image
# plt显示彩色图片
def plt_show0(img):
b,g,r = cv2.split(img)
img = cv2.merge([r, g, b])
plt.imshow(img)
plt.show()
# plt显示灰度图片
def plt_show(img):
plt.imshow(img,cmap='gray')code>
plt.show()
# 图像去噪灰度处理
def gray_guss(image):
image = cv2.GaussianBlur(image, (3, 3), 0)
gray_image = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
return gray_image
# 读取待检测图片
origin_image = cv2.imread('D:/image/car3.jpg')
# 复制一张图片,在复制图上进行图像操作,保留原图
image = origin_image.copy()
# 图像去噪灰度处理
gray_image = gray_guss(image)
# x方向上的边缘检测(增强边缘信息)
Sobel_x = cv2.Sobel(gray_image, cv2.CV_16S, 1, 0)
absX = cv2.convertScaleAbs(Sobel_x)
image = absX
# 图像阈值化操作——获得二值化图
ret, image = cv2.threshold(image, 0, 255, cv2.THRESH_OTSU)
# 显示灰度图像
plt_show(image)
# 形态学(从图像中提取对表达和描绘区域形状有意义的图像分量)——闭操作
kernelX = cv2.getStructuringElement(cv2.MORPH_RECT, (30, 10))
image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernelX,iterations = 1)
# 显示灰度图像
plt_show(image)
# 腐蚀(erode)和膨胀(dilate)
kernelX = cv2.getStructuringElement(cv2.MORPH_RECT, (50, 1))
kernelY = cv2.getStructuringElement(cv2.MORPH_RECT, (1, 20))
#x方向进行闭操作(抑制暗细节)
image = cv2.dilate(image, kernelX)
image = cv2.erode(image, kernelX)
#y方向的开操作
image = cv2.erode(image, kernelY)
image = cv2.dilate(image, kernelY)
# 中值滤波(去噪)
image = cv2.medianBlur(image, 21)
# 显示灰度图像
plt_show(image)
# 获得轮廓
contours, hierarchy = cv2.findContours(image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
for item in contours:
rect = cv2.boundingRect(item)
x = rect[0]
y = rect[1]
weight = rect[2]
height = rect[3]
# 根据轮廓的形状特点,确定车牌的轮廓位置并截取图像
if (weight > (height * 3)) and (weight < (height * 4.5)):
image = origin_image[y:y + height, x:x + weight]
plt_show(image)
#车牌字符分割
# 图像去噪灰度处理
gray_image = gray_guss(image)
# 图像阈值化操作——获得二值化图
ret, image = cv2.threshold(gray_image, 0, 255, cv2.THRESH_OTSU)
plt_show(image)
#膨胀操作
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (4, 4))
image = cv2.dilate(image, kernel)
plt_show(image)
# 查找轮廓
contours, hierarchy = cv2.findContours(image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
words = []
word_images = []
#对所有轮廓逐一操作
for item in contours:
word = []
rect = cv2.boundingRect(item)
x = rect[0]
y = rect[1]
weight = rect[2]
height = rect[3]
word.append(x)
word.append(y)
word.append(weight)
word.append(height)
words.append(word)
# 排序,车牌号有顺序。words是一个嵌套列表
words = sorted(words,key=lambda s:s[0],reverse=False)
i = 0
#word中存放轮廓的起始点和宽高
for word in words:
# 筛选字符的轮廓
if (word[3] > (word[2] * 1.5)) and (word[3] < (word[2] * 5.5)) and (word[2] > 10):
i = i+1
if word[2] < 15:
splite_image = image[word[1]:word[1] + word[3], word[0]-word[2]:word[0] + word[2]*2]
else:
splite_image = image[word[1]:word[1] + word[3], word[0]:word[0] + word[2]]
word_images.append(splite_image)
print(i)
print(words)
for i,j in enumerate(word_images):
plt.subplot(1,7,i+1)
plt.imshow(word_images[i],cmap='gray')code>
plt.show()
#模版匹配
# 准备模板(template[0-9]为数字模板;)
template = ['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',
'藏','川','鄂','甘','赣','贵','桂','黑','沪','吉','冀','津','晋','京','辽','鲁','蒙','闽','宁',
'青','琼','陕','苏','皖','湘','新','渝','豫','粤','云','浙']
# 读取一个文件夹下的所有图片,输入参数是文件名,返回模板文件地址列表
def read_directory(directory_name):
referImg_list = []
for filename in os.listdir(directory_name):
referImg_list.append(directory_name + "/" + filename)
return referImg_list
# 获得中文模板列表(只匹配车牌的第一个字符)
def get_chinese_words_list():
chinese_words_list = []
for i in range(34,64):
#将模板存放在字典中
c_word = read_directory('D:/refer1/'+ template[i])
chinese_words_list.append(c_word)
return chinese_words_list
chinese_words_list = get_chinese_words_list()
# 获得英文模板列表(只匹配车牌的第二个字符)
def get_eng_words_list():
eng_words_list = []
for i in range(10,34):
e_word = read_directory('D:/refer1/'+ template[i])
eng_words_list.append(e_word)
return eng_words_list
eng_words_list = get_eng_words_list()
# 获得英文和数字模板列表(匹配车牌后面的字符)
def get_eng_num_words_list():
eng_num_words_list = []
for i in range(0,34):
word = read_directory('D:/refer1/'+ template[i])
eng_num_words_list.append(word)
return eng_num_words_list
eng_num_words_list = get_eng_num_words_list()
# 读取一个模板地址与图片进行匹配,返回得分
def template_score(template,image):
#将模板进行格式转换
template_img=cv2.imdecode(np.fromfile(template,dtype=np.uint8),1)
template_img = cv2.cvtColor(template_img, cv2.COLOR_RGB2GRAY)
#模板图像阈值化处理——获得黑白图
ret, template_img = cv2.threshold(template_img, 0, 255, cv2.THRESH_OTSU)
# height, width = template_img.shape
# image_ = image.copy()
# image_ = cv2.resize(image_, (width, height))
image_ = image.copy()
#获得待检测图片的尺寸
height, width = image_.shape
# 将模板resize至与图像一样大小
template_img = cv2.resize(template_img, (width, height))
# 模板匹配,返回匹配得分
result = cv2.matchTemplate(image_, template_img, cv2.TM_CCOEFF)
return result[0][0]
# 对分割得到的字符逐一匹配
def template_matching(word_images):
results = []
for index,word_image in enumerate(word_images):
if index==0:
best_score = []
for chinese_words in chinese_words_list:
score = []
for chinese_word in chinese_words:
result = template_score(chinese_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[34+i])
r = template[34+i]
results.append(r)
continue
if index==1:
best_score = []
for eng_word_list in eng_words_list:
score = []
for eng_word in eng_word_list:
result = template_score(eng_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[10+i])
r = template[10+i]
results.append(r)
continue
else:
best_score = []
for eng_num_word_list in eng_num_words_list:
score = []
for eng_num_word in eng_num_word_list:
result = template_score(eng_num_word,word_image)
score.append(result)
best_score.append(max(score))
i = best_score.index(max(best_score))
# print(template[i])
r = template[i]
results.append(r)
continue
return results
word_images_ = word_images.copy()
# 调用函数获得结果
result = template_matching(word_images_)
print(result)
# "".join(result)函数将列表转换为拼接好的字符串,方便结果显示
print( "".join(result))
height,weight = origin_image.shape[0:2]
print(height)
print(weight)
image_1 = origin_image.copy()
cv2.rectangle(image_1, (int(0.2*weight), int(0.75*height)), (int(weight*0.9), int(height*0.95)), (0, 255, 0), 5)
#设置需要显示的字体
fontpath = "font/simsun.ttc"
font = ImageFont.truetype(fontpath,64)
img_pil = Image.fromarray(image_1)
draw = ImageDraw.Draw(img_pil)
#绘制文字信息
draw.text((int(0.2*weight)+25, int(0.75*height)), "".join(result), font = font, fill = (255, 255, 0))
bk_img = np.array(img_pil)
print(result)
print( "".join(result))
plt_show0(bk_img)
声明
本文内容仅代表作者观点,或转载于其他网站,本站不以此文作为商业用途
如有涉及侵权,请联系本站进行删除
转载本站原创文章,请注明来源及作者。