def extract_text(self, is_captcha=False):
image = (self.image if not isinstance(self.image, str) else Image.open(
self.image)).convert('L')
return is_captcha and sub('[\W]', '',
(image_to_string(image)
or '').strip()) or image_to_string(
image) # noqa
def get_text_from_file(self, filepath):
from pyocr import tesseract, builders
from PIL import Image
image = Image.open(filepath)
text = tesseract.image_to_string(image=image, builder=builders.TextBuilder())
return text.strip('\r\n')
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_box_file = "tests/tesseract/" + expected_box_file
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file), lang=lang,
builder=self.builder)
boxes.sort()
self.assertTrue(len(boxes) > 0)
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
try:
# python 2.7
self.assertEqual(type(expected_boxes[i].content), unicode)
self.assertEqual(type(boxes[i].content), unicode)
except NameError:
# python 3
self.assertEqual(type(expected_boxes[i].content), str)
self.assertEqual(type(boxes[i].content), str)
self.assertEqual(boxes[i], expected_boxes[i])
def test_text(image_file, lang='eng'):
print image_file
return tesseract.image_to_string(
Image.open(image_file),
lang=lang,
builder=tesseract.DigitBuilder())
def _read_from_img(self, image_path, lang=None):
boxes = tesseract.image_to_string(Image.open(image_path),
lang=lang,
builder=self._builder)
boxes.sort()
return boxes
def get_digit(self, image, min_score):
if self.recognize_digit:
import pyocr
import pyocr.tesseract as tess
import pyocr.builders
from PIL import Image
digit_ratio = (0.1016, 0.1836, 0.2228, 0.2268)
digits = selectROI(image, ratio=digit_ratio, round8=False)
bw_img = digits
cv2.imwrite('digit_sample.jpg', bw_img)
txt = tess.image_to_string(Image.fromarray(bw_img),
lang='eng',
builder=pyocr.builders.TextBuilder())
txt = txt.replace(',', '').replace('.', '')
if txt.isdigit(
) and int(txt) >= min_score: # reward should not decrease
score = int(txt) / 100.0
else:
score = min_score
# print("TXT:", txt, 'score:', score , 'last_score:', last_score)
# scale score
return score
else:
return 0
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_box_file = "tests/tesseract/" + expected_box_file
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file),
lang=lang,
builder=self.builder)
boxes.sort()
self.assertTrue(len(boxes) > 0)
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
try:
# python 2.7
self.assertEqual(type(expected_boxes[i].content), unicode)
self.assertEqual(type(boxes[i].content), unicode)
except NameError:
# python 3
self.assertEqual(type(expected_boxes[i].content), str)
self.assertEqual(type(boxes[i].content), str)
self.assertEqual(boxes[i], expected_boxes[i])
def test_text_error_file(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "")
copen.side_effect = Exception("Unknown error")
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
with self.assertRaises(Exception):
tesseract.image_to_string(self.image, builder=self.builder)
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def test_text_cannot_open_file(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "")
copen.side_effect = PermissionError(errno.EPERM, "Error opening file")
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
with self.assertRaises(PermissionError):
tesseract.image_to_string(self.image, builder=self.builder)
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def main(self, text_img_name):
txt = tool.image_to_string(
Im.open(text_img_name), lang=self.lang,
builder=pyocr.builders.TextBuilder()
)
return txt
def _read_from_img(self, image_path, lang=None):
boxes = tesseract.image_to_string(
Image.open(image_path),
lang=lang,
builder=self._builder
)
boxes.sort()
return boxes
def find_secret_rects(image, secret_res, lang, tesseract_configs=None):
"""
Find secret rects in an image.
param: Image image
param: list secret_res
param: str lang
param: str tesseract_configs
return: list of rects
rtype: list
"""
# When using pyocr, an input image is converted to RGB (not RGBA).
# During the conversion, transparent pixels are converted into BLACK.
# Sometimes the black pixels get in the way of recognizing text.
#
# For example, macOS's screenshot image taken by Command+Shift+4+Space
# has transparent pixels around an window. This results in a black and
# thick border in the edge of image. The border worsen quality of OCR
# of text near by the border. To avoid this, convert transparent pixels
# into WHITE by pasting image into an white background.
#
# See: http://stackoverflow.com/questions/9166400/convert-rgba-png-to-rgb-with-pil
if image.mode == 'RGBA':
background = Image.new('RGB', image.size, (255, 255, 255))
background.paste(image, mask=image.split()[3]) # Paste image masked by alpha channel [3]
image = background
# offset = (0, 150)
# cropped_image = image.crop((offset[0], offset[1], image.size[0], 220))
offset = (0, 0)
cropped_image = image
builder = ModifiedCharBoxBuilder(cropped_image.size[1])
if tesseract_configs:
builder.tesseract_configs = tesseract_configs
boxes = image_to_string(cropped_image, lang=lang, builder=builder)
if os.environ.get('DEBUG'):
for box in boxes:
print(box.content, box.position)
content = ''.join(box.content for box in boxes)
assert len(boxes) == len(content)
secret_rects = []
for secret_re in secret_res:
for m in secret_re.finditer(content):
matched_boxes = boxes[m.start():m.end()]
matched_rects = [b.position for b in matched_boxes]
for rect in bounding_boxes_by_line(matched_rects):
rect = offset_rect(offset, padding_box(rect, 2))
secret_rects.append(rect)
return secret_rects
def test_char_error(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (1, "Error")
copen.return_value = StringIO(self._get_file_content("boxes"))
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
with self.assertRaises(tesseract.TesseractError) as te:
tesseract.image_to_string(self.image, builder=self.builder)
self.assertEqual(te.exception.status, 1)
self.assertEqual(te.exception.message, "Error")
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def test_char_no_output(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "No file output")
copen.side_effect = FileNotFoundError(
errno.ENOENT, "[Errno 2] No such file or directory: 'output'")
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
with self.assertRaises(tesseract.TesseractError) as te:
tesseract.image_to_string(self.image, builder=self.builder)
self.assertEqual(te.exception.status, -1)
self.assertIn("Unable to find output file (tested",
te.exception.message)
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def __test_txt(self, image_file, expected_output_file, lang="eng"):
image_file = "tests/data/" + image_file
expected_output_file = "tests/tesseract/" + expected_output_file
expected_output = ""
with codecs.open(expected_output_file, "r", encoding="utf-8") as file_descriptor:
for line in file_descriptor:
expected_output += line
expected_output = expected_output.strip()
output = tesseract.image_to_string(Image.open(image_file), lang=lang)
self.assertEqual(output, expected_output)
def main(self, path):
count = 0
other_files = 0
for f in os.listdir(path): #Return list of files in path directory
ext = os.path.splitext(
f
)[1] #Split the pathname path into a pair i.e take .png/ .jpg etc
if ext.lower(
) not in VALIDITY: #Convert to lowercase and check in validity list
other_files += 1 #Increment if other than validity extension found
#sys.stdout.write("Extension other than image is not supported. \n")
continue
else:
count += 1
image_file_name = path + '/' + f #Full /dir/path/filename.extension
txt = tool.image_to_string(
Im.open(image_file_name),
lang=self.lang,
builder=pyocr.builders.TextBuilder())
initial = txt.replace('\a', ' ').replace('\b', ' ').replace(
'\f', ' '
).replace('\n', ' ').replace('\r', '').replace(
'\t', ' '
).replace(
'\v', ' '
) #.replace(' ','_') #.replace('.','_') #Replace \n and \t with space
initial = initial[:60] #Take 1st 100 words
print('Filename:' + initial + '\n')
os.chmod(path, 0o777)
os.rename(image_file_name, path + '/' + initial + ext)
print(
str(count) + (" file" if count == 1 else " files") +
" processed")
if count + other_files == 0:
print("No files found") #No files found
else:
print(
str(count) + " / " + str(count + other_files) +
" files converted")
def __test_txt(self, image_file, expected_output_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_output_file = "tests/tesseract/" + expected_output_file
expected_output = ""
with codecs.open(expected_output_file, 'r', encoding='utf-8') \
as file_descriptor:
for line in file_descriptor:
expected_output += line
expected_output = expected_output.strip()
output = tesseract.image_to_string(Image.open(image_file), lang=lang)
self.assertEqual(output, expected_output)
def __test_text(self, image_file, expected_output_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_output_file = "tests/tesseract/" + expected_output_file
expected_output = ""
with codecs.open(expected_output_file, 'r', encoding='utf-8') \
as file_descriptor:
for line in file_descriptor:
expected_output += line
expected_output = expected_output.strip()
output = tesseract.image_to_string(Image.open(image_file), lang=lang, builder=self.builder)
self.assertEqual(output, expected_output)
def __test_txt(self, image_file, expected_box_file, lang="eng"):
image_file = "tests/data/" + image_file
expected_box_file = "tests/tesseract/" + expected_box_file
with codecs.open(expected_box_file, "r", encoding="utf-8") as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file), lang=lang, builder=self.builder)
boxes.sort()
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
self.assertEqual(boxes[i], expected_boxes[i])
def __test_txt(self, image_file, expected_output_file, lang='eng'):
image_file = os.path.join("tests", "input", "specific", image_file)
expected_output_file = os.path.join("tests", "output", "specific",
"tesseract", expected_output_file)
expected_output = ""
with codecs.open(expected_output_file, 'r', encoding='utf-8') \
as file_descriptor:
for line in file_descriptor:
expected_output += line
expected_output = expected_output.strip()
output = tesseract.image_to_string(Image.open(image_file), lang=lang)
self.assertEqual(output, expected_output)
def rec_img(img):
width = img.size[0]
height = img.size[1]
#构造指数的位置
rangle = (24.5+6.1*length+5,int(height/2),int(width),int(height)) #左、上、右、下
# 打开截图切割
img = img.crop(rangle)
# 将图片放大
(x, y) = img.size
x_s = int(x*2.4)
y_s = int(y*2.4)
imgzoom = img.resize((x_s,y_s),Image.ANTIALIAS)
code = tesseract.image_to_string(imgzoom)
result = re.sub("\D", "", code)
return result
def __test_txt(self, image_file, expected_output_file, lang='eng'):
image_file = os.path.join("tests", "input", "specific", image_file)
expected_output_file = os.path.join(
"tests", "output", "specific", "tesseract", expected_output_file
)
expected_output = ""
with codecs.open(expected_output_file, 'r', encoding='utf-8') \
as file_descriptor:
for line in file_descriptor:
expected_output += line
expected_output = expected_output.strip()
output = tesseract.image_to_string(Image.open(image_file), lang=lang)
self.assertEqual(output, expected_output)
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_box_file = "tests/tesseract/" + expected_box_file
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file), lang=lang,
builder=self.builder)
boxes.sort()
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
self.assertEqual(boxes[i], expected_boxes[i])
def login(username,password):
while 1 :
name="code.jpg"
threshold = 140
table = []
for i in range(256):
if i < threshold:
table.append(0)
else:
table.append(1)
try :
image = requests.get('http://www.ourui.com/userself/RndCode.asp?rndtype=LOGIN_RndCode')
#print(image.content)
f=open(name,"wb+")
f.write(image.content)
f.close()
imagefile = Image.open(name)
#转化到亮度
imgry = imagefile.convert('L')
imgry.save('g'+name)
#二值化
#out = imgry.point(table,'1')
out = imgry.point(lambda x: 255 if x > 141 else 0)
out.save('b'+name)
#识别
val=tesseract.image_to_string(out)
print(val)
imgreq=requests.get("http://www.ourui.com/UserSelf/rndcode_check.asp?name=lg&rndcode="+val)
print(imgreq.content)
if imgreq.content=="0" or imgreq.content=="":
continue
r = requests.post("http://www.ourui.com/userself/login.asp", data={"username":username,"userpass":password})
r.content
print("login ok register",self.domain)
break
except:
time.sleep(1)
print("矮油网速太渣，登录失败，重新登录！")
def get_data(self, region, rerun=0,
image_optimizer=None, data_optimizer=None, threshold=None):
'''Gets data out of a single region
:param rerun: incrementing integer for every rerun.
:param image_optimizer: callback, takes PIL.Image and rerun as params.
:param data_optimizer: callback, takes String and rerun as params.
:param threshold: callback, takes String and should return False for a rerun.
'''
if image_optimizer:
optiregion = image_optimizer(region, rerun)
data = tesseract.image_to_string(optiregion)
if data_optimizer:
data = data_optimizer(data, rerun)
if threshold and not threshold(data):
return self.get_data(
region, rerun + 1,
image_optimizer, data_optimizer, threshold)
else:
return data
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = os.path.join("tests", "input", "specific", image_file)
expected_box_file = os.path.join("tests", "output", "specific",
"tesseract", expected_box_file)
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file),
lang=lang,
builder=self.builder)
boxes.sort()
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
self.assertEqual(boxes[i], expected_boxes[i])
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = os.path.join("tests", "input", "specific", image_file)
expected_box_file = os.path.join(
"tests", "output", "specific", "tesseract", expected_box_file
)
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
boxes = tesseract.image_to_string(Image.open(image_file), lang=lang,
builder=self.builder)
boxes.sort()
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
self.assertEqual(boxes[i], expected_boxes[i])
def test_text(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "")
copen.return_value = StringIO(self._get_file_content("text"))
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
result = tesseract.image_to_string(self.image,
builder=self.builder)
self.assertEqual(result, self._get_file_content("text").strip())
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def test_write_read(self):
original_boxes = tesseract.image_to_string(Image.open("tests/data/test.png"), builder=self.builder)
self.assertTrue(len(original_boxes) > 0)
(file_descriptor, tmp_path) = tempfile.mkstemp()
try:
# we must open the file with codecs.open() for utf-8 support
os.close(file_descriptor)
with codecs.open(tmp_path, "w", encoding="utf-8") as fdescriptor:
self.builder.write_file(fdescriptor, original_boxes)
with codecs.open(tmp_path, "r", encoding="utf-8") as fdescriptor:
new_boxes = self.builder.read_file(fdescriptor)
self.assertEqual(len(new_boxes), len(original_boxes))
for i in range(0, len(original_boxes)):
self.assertEqual(new_boxes[i], original_boxes[i])
finally:
os.remove(tmp_path)
def test_char(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "")
copen.return_value = StringIO(self._get_file_content("boxes"))
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
result = tesseract.image_to_string(self.image,
builder=self.builder)
for box in result:
self.assertIsInstance(box, builders.Box)
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def __test_txt(self, image_file, expected_box_file, lang='eng'):
image_file = "tests/data/" + image_file
expected_box_file = "tests/tesseract/" + expected_box_file
boxes = tesseract.image_to_string(Image.open(image_file), lang=lang,
builder=self.builder)
boxes.sort()
with codecs.open(expected_box_file, 'r', encoding='utf-8') \
as file_descriptor:
expected_boxes = self.builder.read_file(file_descriptor)
expected_boxes.sort()
self.assertEqual(len(boxes), len(expected_boxes))
for i in range(0, min(len(boxes), len(expected_boxes))):
for j in range(0, len(boxes[i].word_boxes)):
self.assertEqual(type(boxes[i].word_boxes[j]),
type(expected_boxes[i].word_boxes[j]))
self.assertEqual(boxes[i], expected_boxes[i])
def test_write_read(self):
original_boxes = tesseract.image_to_string(
Image.open("tests/data/test.png"), builder=self.builder)
self.assertTrue(len(original_boxes) > 0)
(file_descriptor, tmp_path) = tempfile.mkstemp()
try:
# we must open the file with codecs.open() for utf-8 support
os.close(file_descriptor)
with codecs.open(tmp_path, 'w', encoding='utf-8') as fdescriptor:
self.builder.write_file(fdescriptor, original_boxes)
with codecs.open(tmp_path, 'r', encoding='utf-8') as fdescriptor:
new_boxes = self.builder.read_file(fdescriptor)
self.assertEqual(len(new_boxes), len(original_boxes))
for i in range(0, len(original_boxes)):
self.assertEqual(new_boxes[i], original_boxes[i])
finally:
os.remove(tmp_path)
def test_text_non_rgb_image(self, run_tesseract, copen, temp_dir):
"""This tests that image_to_string works with non RGB mode images and
that image is converted in function."""
image = self.image.convert("L")
run_tesseract.return_value = (0, "")
copen.return_value = StringIO(self._get_file_content("text"))
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
result = tesseract.image_to_string(image, builder=self.builder)
self.assertEqual(result, self._get_file_content("text").strip())
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def get_data(self,
region,
rerun=0,
image_optimizer=None,
data_optimizer=None,
threshold=None):
'''Gets data out of a single region
:param rerun: incrementing integer for every rerun.
:param image_optimizer: callback, takes PIL.Image and rerun as params.
:param data_optimizer: callback, takes String and rerun as params.
:param threshold: callback, takes String and should return False for a rerun.
'''
if image_optimizer:
optiregion = image_optimizer(region, rerun)
data = tesseract.image_to_string(optiregion)
if data_optimizer:
data = data_optimizer(data, rerun)
if threshold and not threshold(data):
return self.get_data(region, rerun + 1, image_optimizer,
data_optimizer, threshold)
else:
return data
def test_digits(self, run_tesseract, copen, temp_dir):
run_tesseract.return_value = (0, "")
copen.return_value = StringIO(self._get_file_content("digits"))
with TemporaryDirectory(prefix="tess_") as tmpdir:
enter = MagicMock()
enter.__enter__.return_value = tmpdir
temp_dir.return_value = enter
with open(os.path.join(tmpdir, "output.txt"), "w") as fh:
fh.write("")
result = tesseract.image_to_string(self.image,
builder=self.builder)
for digit in result:
self.assertIsInstance(int(digit), int)
run_tesseract.assert_called_once_with(
"input.bmp",
"output",
cwd=tmpdir,
lang=None,
flags=self.builder.tesseract_flags,
configs=self.builder.tesseract_configs,
)
def get_text(img_byte):
"""识别"""
im = Image.open(BytesIO(img_byte))
# 去底色
im = im.point(lambda i: 255 if i > 180 else 0)
# 去除干扰线
size = im.size
pimx = im.load()
for x in range(size[0]):
for y in range(size[1]):
px = pimx[x, y]
if px[0] == 0 and px[1] == 0 and px[2] == 0:
pimx[x, y] = pimx[x, 0 if y == 0 else y - 1]
im = ImageOps.invert(im).convert("1")
old_width, old_height = im.size
im.thumbnail((old_width*0.7, old_height*0.7))
arr = np.array(im).sum(axis=0)
print(arr)
# 剪裁
region = im.crop((36, 1, 105, 34))
# OCR
builder = tesseract.builders.DigitBuilder()
digits_address = os.path.join(os.getcwd(), 'config/digits')
print("digits_address -> ", digits_address)
builder.tesseract_configs = ['-psm', '7', digits_address]
result = tesseract.image_to_string(region, 'eng', builder)
code_text = result.replace(' ', '')
print("out text -> ", code_text)
if len(code_text) == 4:
return str(code_text)
else:
return "0"
def get_text(img_byte, card_modul):
"""识别卡号"""
im = Image.open(BytesIO(img_byte))
print(card_modul['typeId'], card_modul['name'])
if card_modul['typeId'] == 1:
# 剪裁
region = im.crop(card_modul["cut"])
elif card_modul['typeId'] == 2:
# 剪裁
region = im.crop(card_modul["cut"])
elif card_modul['typeId'] == 3:
# 剪裁
region = im.crop(card_modul["cut"])
region = ImageOps.invert(region).convert("L")
region.show()
# OCR
builder = tesseract.builders.DigitBuilder()
digits_address = os.path.join(os.getcwd(), 'config/digits')
builder.tesseract_configs = ['-psm', '7', digits_address]
result = tesseract.image_to_string(region, 'eng', builder)
return result.replace(' ', '')
def extract(image_file, spellchecker=None):
text = pytesseract.image_to_string(Image.open(image_file), lang="deu")
clean_text = []
for word in re.findall(r"\w+", text):
word = word.strip()
if not word:
continue
if word.isdigit():
clean_text.append(word)
continue
if len(word) == 1:
clean_text.append(word)
continue
if spellchecker:
correction = spellchecker(word.lower())
if correction and not isinstance(correction, str):
# some spellcheckers return a list of suggestions -> use
# first suggestion
correction = correction[0]
else:
correction = None
if correction:
if word[0].isupper():
# keep capitalization of first char.
correction = correction[0].upper() + correction[1:]
clean_text.append(correction)
else:
clean_text.append(word + "?")
emails = []
for email in re.findall(
"\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}\b",
text,
flags=re.UNICODE + re.IGNORECASE,
):
_log.debug("email: %s", email)
emails.append(email)
return " ".join(clean_text), emails
from pyocr import tesseract
from PIL import Image
imagefile = Image.open('code.jpg')
val = tesseract.image_to_string(imagefile)
print(val)
def scanImage(self, img):
if tesseract.is_available():
return tesseract.image_to_string(Image.open(img))
else:
return "FAIL"
def getText(filename):
img = Image.open(filename)
text = tesseract.image_to_string(img)
return text
from PIL import Image
from pyocr import tesseract
pic_list = ['pic1.png', 'pic2.png']
for i in pic_list:
im = Image.open(i)
im = im.convert('L') # 图片转换为灰色图像
# 保存转换后的图片
im.save("temp.png")
code = tesseract.image_to_string(im)
print(code)
import pytesseract as tess
import pytesseract as Output
from PIL import Image
import cv2 as cv
import re
#line 7 is optional
tess.pytesseract.tesseract_cmd = r'C:\Program Files (x86)\Tesseract-OCR\tesseract.exe'
from pyocr.tesseract import image_to_string
text = image_to_string(
Image.open(r'C:\Users\Admin\projects\webapp\media\report\images\med1.png'),
lang='eng')
img = cv.cv2.imread(
r'C:\Users\Admin\projects\webapp\media\report\images\med1.png')
custom_config = r'-c tessedit_char_whitelist=medical --psm 6'
print(tess.pytesseract.image_to_string(img, config=custom_config))
#TESSERACT_CMD = 'tesseract.exe' if os.name == 'nt' else 'tesseract'
#m = Image.open('c:/Users/Admin/projects/webapp/media/report/images/1.png')
#text = pytesseract.image_to_string('m')
import numpy as np
import cv2
from PIL import Image
import sys
import pyocr.tesseract as tess
import pyocr.builders
import time
import os, subprocess
ts = time.time()
txt = tess.image_to_string(Image.open('digit_sample.jpg'),
lang='eng',
builder=pyocr.builders.TextBuilder())
#print(txt)
#print(time.time() - ts)
# txt is a Python string
def image_to_string(img, cleanup=True, plus=''):
# cleanup为True则识别完成后删除生成的文本文件
# plus参数为给tesseract的附加高级参数
# subprocess.check_output('tesseract ' + img + ' ' +
# img + ' ' + plus, shell=True) # 生成同名txt文件
os.popen('tesseract ' + img + ' ' + img + ' ' + plus)
text = ''
with open(img + '.txt', 'r') as f:
text = f.read().strip()
from pyocr import builders
from PIL import Image, ImageEnhance, ImageFilter
def test_text(image_file, lang='eng'):
print image_file
return tesseract.image_to_string(
Image.open(image_file),
lang=lang,
builder=tesseract.DigitBuilder())
print test_text('./123.png')
print Image.open('./123.png')
print tesseract.image_to_string(Image.open('./11.jpg'),
lang='eng')
image_name = "./123.png"
im = Image.open(image_name)
im = im.filter(ImageFilter.MedianFilter())
enhancer = ImageEnhance.Contrast(im)
im = enhancer.enhance(2)
im = im.convert('1')
#im.show()
#all by pixel
s = 12 #start postion of first number
w = 10 #width of each number
h = 15 #end postion from top
t = 2 #start postion of top
im_new = []
#split four numbers in the picture
except ValueError as e:
continue
# Use wand to convert each page in the PDF into an image blob
# Loop over blobs
for img in image_jpeg.sequence:
img_page = Image(image=img)
# append as a blob into the req_image list
req_image.append(img_page.make_blob('jpeg'))
# run OCR over the image blobs
txt = ""
for img in req_image:
try:
txt += tool.image_to_string(
PI.open(io.BytesIO(img)),
lang=lang,
builder=pyocr.builders.TextBuilder())
error_flag = False
except OSError as e:
error_flag = True
print("#####################")
print("Error: txt")
print("#####################")
final_text.append(txt)
if error_flag:
continue
print("#####################")
print("TEXT FOUND:")
print(final_text)
