class dyOCR():
def __init__(self):
from cnocr import CnOcr
self.ocr = CnOcr()
def resFromFilePath(self, fp):
get = self.ocr.ocr(fp)
if get == []:
return "-1 暂时不能识别"
get = "".join(get[0])
get = get.replace("o", "0")
get = get.replace("O", "0")
return get
def resFromB64(self, b64):
img_data = base64.b64decode(b64)
# nparr = np.fromstring(img_data, np.uint8)
nparr = np.frombuffer(img_data, np.uint8)
img_np = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
get = self.ocr.ocr(img_np)
if get == []:
return "-1 暂时不能识别"
get = "".join(get[0])
get = get.replace("o", "0")
get = get.replace("O", "0")
return get
def test_cand_alphabet2():
img_fp = os.path.join(example_dir, 'hybrid.png')
ocr = CnOcr(name='instance1')
pred = ocr.ocr(img_fp)
pred = [''.join(line_p) for line_p in pred]
print("Predicted Chars:", pred)
assert len(pred) == 1 and pred[0] == 'o12345678'
ocr.set_cand_alphabet(NUMBERS)
pred = ocr.ocr(img_fp)
pred = [''.join(line_p) for line_p in pred]
print("Predicted Chars:", pred)
assert len(pred) == 1 and pred[0] == '012345678'
def test_multiple_instances(model_name):
global INSTANCE_ID
print('test multiple instances for model_name: %s' % model_name)
img_fp = os.path.join(example_dir, 'hybrid.png')
INSTANCE_ID += 1
print('instance id: %d' % INSTANCE_ID)
cnocr1 = CnOcr(model_name, name='instance-%d' % INSTANCE_ID)
print_preds(cnocr1.ocr(img_fp))
INSTANCE_ID += 1
print('instance id: %d' % INSTANCE_ID)
cnocr2 = CnOcr(model_name,
name='instance-%d' % INSTANCE_ID,
cand_alphabet=NUMBERS)
print_preds(cnocr2.ocr(img_fp))
def getOCR(img):
"""使用CnOcr识别图片内容，按行输出结果至res列表的函数
"""
global result_str, result
buffer = QtCore.QBuffer()
buffer.open(QtCore.QBuffer.ReadWrite)
img.save(buffer, "PNG")
pil_img = Image.open(io.BytesIO(buffer.data()))
image_array = np.array(pil_img)
buffer.close()
ocr = CnOcr()
res = ocr.ocr(image_array)
try:
# 拼接字符串并保存至results列表
result_str = []
for iter_item in res:
result_str.append(''.join(iter_item))
result = "\n".join(result_str)
except Exception as e:
print(f"INFO: Unable to read text from image, did not copy")
notify(f"Unable to read text from image, did not copy")
if result:
pyperclip.copy(result)
print(f'INFO: Copied "{result}" to the clipboard')
notify(f'Copied "{result}" to the clipboard')
else:
print(f"INFO: Unable to read text from image, did not copy")
notify(f"Unable to read text from image, did not copy")
def pathCallBack():
# , ('All Files', '*')
filePath = filedialog.askopenfilename(title='Select picture to OCR', filetypes=[('PNG', '*.png'), ('JPG', '*.jpg'),('JPEG', '*.jpeg'), ('bmp', '*.bmp')])
if (filePath != ''):
strPath.set(filePath)
ocr = CnOcr()
img_fp = ''
# img_fp = 'D://download/multi-line_cn1.png'
img_fp = filePath
img = mx.image.imread(img_fp, 1)
res = ocr.ocr(img)
# print("Predicted Chars:", res)
# print(type(res))
strResult = ""
for s in res:
strResult = strResult + ''.join('%s' % id for id in s) + "\n"
txtResult.delete(0.0, tk.END)
txtResult.insert(tk.INSERT, strResult)
txtResult.update()
def test_cand_alphabet():
from cnocr import NUMBERS
img_fp = os.path.join(example_dir, 'hybrid.png')
ocr = CnOcr()
pred = ocr.ocr(img_fp)
pred = [''.join(line_p) for line_p in pred]
print("Predicted Chars:", pred)
assert len(pred) == 1 and pred[0] == 'o12345678'
ocr = CnOcr(cand_alphabet=NUMBERS)
pred = ocr.ocr(img_fp)
pred = [''.join(line_p) for line_p in pred]
print("Predicted Chars:", pred)
assert len(pred) == 1 and pred[0] == '012345678'
def get_list(self):
std=CnStd()
ocr=CnOcr()
ocr._model_name='conv-lite-fc'
print(ocr._model_name)
ocr_res2=ocr.ocr(img_fp=self.filepath)
box_info_list=std.detect(self.filepath,pse_threshold=0.7,pse_min_area=150,context='gpu',height_border=0.10)
image=Image.open(self.filepath)
fontstyle=ImageFont.truetype('./simhei.ttf',13,encoding='utf-8')
draw=ImageDraw.Draw(image)
for box_info in box_info_list:
print('a')
print('a')
print(box_info)
info_box=box_info['box']
crp_img=box_info['cropped_img']
ocr_res1=ocr.ocr_for_single_line(crp_img)
print('result: %s' % ''.join(str(ocr_res1)))
x1,y1=info_box[0,0],info_box[0,1]
x2,y2=info_box[1,0],info_box[1,1]
x3,y3=info_box[2,0],info_box[2,1]
x4,y4=info_box[3,0],info_box[3,1]
draw.polygon([(x1,y1),(x4,y4),(x3,y3),(x2,y2)],outline=(255,0,0))
draw.text((x4,y4),str(ocr_res1),(200,0,0),font=fontstyle)
image.show()
print(ocr_res2)
return box_info_list
class FTC(object):
# font_to_chinese
def __init__(self, font='HYZiKuTangChangLinTiW-2.ttf'):
self.img_font = ImageFont.truetype(font, 60)
self.fill_color = "#000000"
self.ocr = CnOcr()
if not os.path.isdir('jpg'):
os.mkdir('jpg')
pass
def make_all_font_img(self):
img = Image.new('RGB', (70 * 20, 80), color='white')
draw = ImageDraw.Draw(img)
draw.text((0, 0), u'abcdefghigklmnopqrstuvwxyz', font=self.img_font, fill=self.fill_color)
img.show()
img.save('jpg/1.jpg')
def make_img(self):
img = Image.new('RGB', (140, 80), color='white')
draw = ImageDraw.Draw(img)
draw.text((0, 0), chr(0xe41b), font=self.img_font, fill=self.fill_color)
# img.show()
# img.save('jpg/1.jpg')
# print(pytesseract.image_to_string(img, lang='chi_sim'))
pass
def ret_key_font(self):
self.make_all_font_img()
res = self.ocr.ocr('jpg/1.jpg')
return res
def main():
parser = argparse.ArgumentParser()
parser.add_argument("--model_name",
help="model name",
type=str,
default='conv-lite-fc')
parser.add_argument("--model_epoch",
type=int,
default=None,
help="model epoch")
parser.add_argument("-f", "--file", help="Path to the image file")
parser.add_argument(
"-s",
"--single-line",
default=False,
help="Whether the image only includes one-line characters",
)
args = parser.parse_args()
ocr = CnOcr(model_name=args.model_name, model_epoch=args.model_epoch)
if args.single_line:
res = ocr.ocr_for_single_line(args.file)
else:
res = ocr.ocr(args.file)
logger.info("Predicted Chars: %s", res)
def OCR():
ocr = CnOcr()
QR_Image = getScreenImage()
#res = ocr.ocr_for_single_line(QR_Image)
res = ocr.ocr(QR_Image)
print("Predicted Chars:", res)
easygui.msgbox(res)
def extract_img(path: str):
ocr = CnOcr()
res = ocr.ocr(path)
lines = []
for obj in res:
line = "".join(obj[0])
lines.append(line)
data = "\n".join(lines)
os_remove(path)
return data
def cnocr_ocr(pic_path):
from cnocr import CnOcr
ocr = CnOcr()
res = ocr.ocr(pic_path)
for i in res[:-1]:
for j in i[0]:
print(j, end='')
print()
for i in res[-1][0]:
print(i, end='')
def cnocr_ocr(pic_path):
from cnocr import CnOcr
ocr = CnOcr()
res = ocr.ocr(pic_path)
for r in res[:-1]:
for c in r:
print(c, end='')
print()
for r in res[-1]:
for c in r:
print(c, end='')
def ocr_image(fname):
print('OCR %s' % fname)
global _cnocr
if _cnocr is None:
_cnocr = CnOcr()
res = _cnocr.ocr(fname)
text = '\n'.join([''.join(t) for t in res])
return text
class OCR_cnocr(object):
"""docstring for OCR_tesseract"""
def __init__(self):
self.ocr = CnOcr()
def run(self, img):
nary = np.asarray(img)
res = self.ocr.ocr(nary)
content = ""
for r in res:
for x in r:
content = content + x
return content
def main():
parser = argparse.ArgumentParser()
parser.add_argument("-f", "--file", help="Path to the image file")
parser.add_argument(
"-s",
"--single-line",
default=False,
help="Whether the image only includes one-line characters")
args = parser.parse_args()
ocr = CnOcr()
if args.single_line:
res = ocr.ocr_for_single_line(args.file)
else:
res = ocr.ocr(args.file)
print("Predicted Chars:", res)
def txtOcr():
if request.method == 'POST':
file = request.files['file']
if file and allowed_file(file.filename):
filename = secure_filename(''.join(lazy_pinyin(file.filename)))
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
file_url = url_for('uploaded_file', filename=filename)
import mxnet as mx
from cnocr import CnOcr
ocr = CnOcr()
img_fp = os.path.join(app.config['UPLOAD_FOLDER'], filename)
img = mx.image.imread(img_fp, 1)
res = ocr.ocr(img)
print("Predicted Chars:", res)
return html + '<br><img src=' + file_url + '><p>预测结果如下：' + str(
res) + '</p>'
return html
def multLineOcr(request):
imgs = request.POST.get("img")
img = imgs.replace(" ", "+")
res = base64.b64decode(img)
# res = base64.urlsafe_b64encode(img)
nparr = np.fromstring(res, np.uint8)
img = cv2.imdecode(nparr, cv2.COLOR_BGR2RGB)
imgb = img
img = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# cv2.imwrite("img.png", imgb)
ocr = CnOcr()
res = ocr.ocr(img)
return HttpResponse(
json.dumps(res)
)
def cn_ocr(self, _range, delay_time=5):
"""
cn ocr识别数字
:param delay_time: ocr识别延迟时间
:param _range: 验证码截图区域坐标(左x,左y,右x,右y)
:return: 识别到的数字
"""
global sms_code
self.get_code_pic(_range)
cn_ocr = CnOcr(model_name="conv-lite-fc",
context="cpu",
root="conv-lite-fc")
ret = cn_ocr.ocr("ios_code_pic.png")
result = ""
for v in ret:
result += "".join(v)
find_all = re.findall(r'\'[\d]{6}\'', str(result))
if len(find_all) != 1:
find_all = re.findall(r'([\d]{6})[\u3002]', str(result))
if len(find_all) != 1:
find_all = re.findall(r'(您的验证码为[\d]{6})', str(result))
# 识别结果
self.logger.info("CnOCR识别结果：" + result)
if len(find_all) == 1:
code = find_all[0].strip("'")
if sms_code == code:
self.logger.info("暂未获取到最新验证码，%d秒后重试" % delay_time)
time.sleep(delay_time)
return self.cn_ocr(_range, delay_time)
else:
sms_code = code
return code
else:
self.logger.info("暂未获取到最新验证码，%d秒后重试" % delay_time)
time.sleep(delay_time)
return self.cn_ocr(_range, delay_time)
class FTC(object):
# font_to_chinese
def __init__(self, font='shaoq.woff'):
self.img_font = ImageFont.truetype(font, 60)
self.fill_color = "#000000"
self.font = TTFont('shaoq.woff')
self.keys = list(key for key in self.font['glyf'].keys()
if key.startswith('uni'))
self.ocr = CnOcr()
if not os.path.isdir('jpg'):
os.mkdir('jpg')
pass
def make_all_font_img(self):
img = Image.new('RGB', (70 * len(self.keys), 80), color='white')
draw = ImageDraw.Draw(img)
draw.text(
(0, 0),
''.join(
chr(int(f'0x{key.strip("uni")}', 16)) for key in self.keys),
font=self.img_font,
fill=self.fill_color)
# img.show()
img.save('jpg/1.jpg')
def make_img(self):
img = Image.new('RGB', (140, 80), color='white')
draw = ImageDraw.Draw(img)
draw.text((0, 0),
chr(0xe41b),
font=self.img_font,
fill=self.fill_color)
# img.show()
# img.save('jpg/1.jpg')
# print(pytesseract.image_to_string(img, lang='chi_sim'))
pass
def ret_key_font(self):
self.make_all_font_img()
res = self.ocr.ocr('jpg/1.jpg')
return dict(zip(self.keys, res[0]))
def POST(self):
x = web.input(myfile={})
filedir = './upload_file' # change this to the directory you want to store the file in.
if 'myfile' in x: # to check if the file-object is created
filepath = x.myfile.filename.replace(
'\\',
'/') # replaces the windows-style slashes with linux ones.
filename = filepath.split(
'/'
)[-1] # splits the and chooses the last part (the filename with extension)
fout = open(
filedir + '/' + filename, 'wb'
) # creates the file where the uploaded file should be stored
fout.write(x.myfile.file.read()
) # writes the uploaded file to the newly created file.
fout.close() # closes the file, upload complete.
myOcr = CnOcr()
resultData = myOcr.ocr(filedir + '/' + filename)
del myOcr
gc.collect()
jsonStr = json.dumps(resultData, cls=NumpyEncoder)
return jsonStr
from cnocr import CnOcr ocr = CnOcr() res = ocr.ocr(r'E:\企业微信截图_15927938103453.png')
import sys
from cnocr import CnOcr
ocr = CnOcr()
res = ocr.ocr(sys.argv[1])
print("Predicted Chars:", res)
from cnocr import CnOcr
import cnocr
from cnstd import CnStd
std=CnStd()
ocr=CnOcr()
box_info_list=std.detect('E:\\Work Place\\pocr\\pic\\2.png')
res=ocr.ocr('E:\\Work Place\\pocr\\pic\\1.png')
for box_info in box_info_list:
crp_img=box_info['cropped_img']
ocr_res=ocr.ocr_for_single_line(crp_img)
print('result: %s' % ''.join(ocr_res))
import mxnet as mx
import sys
import json
from cnocr import CnOcr
class NumpyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj,
(np.int_, np.intc, np.intp, np.int8, np.int16, np.int32,
np.int64, np.uint8, np.uint16, np.uint32, np.uint64)):
return int(obj)
elif isinstance(obj, (np.float_, np.float16, np.float32, np.float64)):
return float(obj)
elif isinstance(obj, (np.ndarray, )): #### This is the fix
return obj.tolist()
return json.JSONEncoder.default(self, obj)
ocr = CnOcr()
img = mx.image.imread(str(sys.argv[1]), 1)
res = ocr.ocr(img)
jsonStr = json.dumps(res, cls=NumpyEncoder)
print(jsonStr)
def conn_USTC(user):
'''通过统一身份认证连接综合教务系统，并保持连接'''
# 使用requests方法实现，https://blog.csdn.net/xc_zhou/article/details/81021496?utm_medium=distribute.pc_relevant_t0.none-task-blog-2%7Edefault%7EBlogCommendFromMachineLearnPai2%7Edefault-1.control&dist_request_id=&depth_1-utm_source=distribute.pc_relevant_t0.none-task-blog-2%7Edefault%7EBlogCommendFromMachineLearnPai2%7Edefault-1.control，https://blog.csdn.net/qq_37616069/article/details/80376776
# https://www.jianshu.com/p/8cd6e9bc2680，先chrome开发者工具分析登录过程，再进行模仿
# 也可考虑使用webdriver实现，见https://blog.csdn.net/Haven200/article/details/103208795
user = user['user']
# 首次访问，获取统一身份认证地址
login_url = 'https://jw.ustc.edu.cn/login'
headers = {
'Accept':
'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9',
'Accept-Encoding':
'gzip, deflate, br',
'Accept-Language':
'zh-CN,zh;q=0.9,en-US;q=0.8,en;q=0.7,ja-JP;q=0.6,ja;q=0.5',
'cache-control':
'max-age=0',
'sec-ch-ua':
'"Google Chrome";v="93", " Not;A Brand";v="99", "Chromium";v="93"',
'sec-ch-ua-mobile':
'?0',
'sec-ch-ua-platform':
'"Windows"',
'sec-fetch-dest':
'document',
'sec-fetch-mode':
'navigate',
'sec-fetch-site':
'none',
'sec-fetch-user':
'******',
'upgrade-insecure-requests':
'1',
'User-Agent':
'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/93.0.4577.82 Safari/537.36'
}
print('connecting to %s ...' % login_url)
res = requests.get(url=login_url, headers=headers)
parser = Parser_loginurl()
parser.feed(res.text)
parser.close()
print(res.text)
print("获取的统一身份认证地址:{0}".format(parser.links[0]))
# 解析并生成下一个url
urlparser = urlparse(login_url)
host = urlparser.netloc
ucaslogin_url = urlparser.scheme + '://' + host + parser.links[0]
# print(ucaslogin_url)
# 访问教务处与统一身份认证的接口，会自动重定向到统一身份认证地址
session = requests.session()
print('connecting to %s ...' % ucaslogin_url)
headers = {
'Accept':
'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9',
'Accept-Encoding':
'gzip, deflate, br',
'Accept-Language':
'zh-CN,zh;q=0.9,en-US;q=0.8,en;q=0.7,ja-JP;q=0.6,ja;q=0.5',
'User-Agent':
'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4389.114 Safari/537.36'
}
res = session.get(url=ucaslogin_url, headers=headers)
# print(res.cookies)
# 获取验证码图片，识别验证码
validate_code = '1234'
## 解析前面获取的html，得到验证码图片链接
parser = Parser_validateimg()
parser.feed(res.text)
parser.close()
print("验证码图片地址:{0}".format(parser.links[0]))
img_url = parser.links[0]
path = './validatecode.jpg'
## 获取并保存验证码图片
res = requests.get(img_url)
with open(path, 'wb') as fp:
fp.write(res.content)
fp.close()
## 识别验证码
ocr = CnOcr()
ocr_res = ocr.ocr(path)
print(res)
# 建立统一身份认证连接，并保持连接
# 登陆时提交表格用到这里的参数，是通过chrome开发者工具查看请求的模式并模仿
headers['Referer'] = res.url
headers['Cookie'] = 'JSESSIONID=' + res.cookies['JSESSIONID']
params = {
'model': 'uplogin.jsp',
'service': '',
'warn': '',
'showCode': '1',
'username': user['username'],
'password': user['password'],
'LT': validate_code,
'button': ''
}
print('connecting to {0},{1},{2} ...'.format(res.url, user['username'],
user['password']))
res = session.post(url=res.url, data=params, headers=headers)
# print(res.headers)
# print(res.url)
# 验证登陆成功
return session
from cnocr import CnOcr
def convert(s):
# initialization of string to ""
str1 = ""
# using join function join the list s by
# separating words by str1
return (str1.join(s))
# Full Screen 2880 x 1920
imgfile = "e:\\bwin\\tmp\\code3.jpg"
ocr = CnOcr()
res = ocr.ocr(imgfile)
txt = [''.join(s) for s in res]
print("Content:\n", txt)
import mxnet as mx
import cv2
ocr = CnOcr(name='img-tarn', model_name='densenet-lite-gru')
"""
example
"""
# response = requests.get('https://image.jiandan100.cn/images/cqaimages/42/255/2817944_q.jpg')
response = requests.get(
'http://8.210.115.9/img.php?num=NjMyMzgy&x=MzM5NDgwNTgzNTk=&s=77600120640')
img_fp = Image.open(BytesIO(response.content))
# img_fp = Image.open('img.png')
np_image = np.array(img_fp)
gray = cv2.cvtColor(np_image, cv2.COLOR_BGR2GRAY)
# print(gray)
t, rst = cv2.threshold(gray, 168, 255, cv2.THRESH_BINARY_INV)
# cv2.imshow('t', gray)
# cv2.imshow('rst', rst)
# cv2.waitKey(0)
# img = mx.image.imread(img_fp, 1).asnumpy()
res = ocr.ocr(rst)
text = '\n'.join([''.join(a) for a in res])
print("Predicted Chars:", text)
#-*- codeing = utf-8 -*-
#@Time : 2020/12/30 下午2:13
#@Author : 江啸栋19262010049
#@File : run.py
#@Software : PyCharm
from cnocr import CnOcr
ocr = CnOcr()
res = ocr.ocr('examples/00010991.jpg')
print("Predicted Chars:", res)
right = 450
bottom = 470
image_obj2 = new_img.crop((left, top, right, bottom))
# image_obj2.show()
#纳税人识别号
left = 155
top = 470
right = 450
bottom = 490
image_text3 = new_img.crop((left, top, right, bottom))
#展示图片
image_text3.show()
# 开票人
left = 528
top = 550
right = 670
bottom = 600
image_obj4 = new_img.crop((left, top, right, bottom))
image_obj4.show()
image_obj2.save("tmp.jpg")
ocr = CnOcr()
res = ocr.ocr("tmp.jpg")
print("".join(res[0]))
image_obj4.save("tmp.jpg")
ocr = CnOcr()
res = ocr.ocr("tmp.jpg")
print("".join(res[0]))
