def OCR(self, plateImage):
tessdata_dir_config = '--tessdata-dir "C:\\Program Files\\Tesseract-OCR\\tessdata"'
cv2.imwrite("a.jpg", plateImage)
textOfImage = tools.image_to_string(Image.open("a.jpg"),
builder=builders.TextBuilder())
#textOfImage = pytesseract.image_to_string(plateImage, lang="eng",config=tessdata_dir_config)
return textOfImage
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 processing():
global rectangle,thresh,erosion_iters,most_common_filter
#
r = cv2.getTrackbarPos('Threshold', 'Inputs') #changed from frame to Inputs
#
cv2.namedWindow('image')
cv2.setMouseCallback('image', draw_shape)
#
img_temp, img, roi = initialize_images()
rectangle = None
scale = 1
factor = 0.75
#
while(1):
sleep(0.2)
img = img_temp.copy()
#
### ROI ###########################################################
if is_rectangle(rectangle):
roi = img_temp[rectangle[0][1]:rectangle[1][1], rectangle[0][0]:rectangle[1][0]]
cv2.rectangle(img,rectangle[0],rectangle[1],(0,255,0),0)
else:
roi = None
### FILTER ########################################################
# retrieving parameters from GUI slidebars
### DISPLAY #######################################################
cv2.imshow('image',img)
if roi is not None: #changed from "if roi != None" which gave array related error
kernel = np.ones((5, 5), np.uint8)
roi = cv2.cvtColor(roi, cv2.COLOR_BGR2GRAY)
roi = cv2.threshold(roi, thresh, 255, cv2.THRESH_BINARY)[1]
roi = cv2.erode(roi, kernel, iterations=erosion_iters)
cv2.imshow('ROI',roi)
if roi is not None and not drawing: #changed from "if roi != None" which gave array related error
### OCR ###########################################################
tool = pyocr.get_available_tools()[0] #
lang = 'letsgodigital'#'letsgodigital'#"eng" #export TESSDATA_PREFIX=/path/to/tessdata/folder
txt = tool.image_to_string(Image.fromarray(roi), lang=lang, builder=builders.TextBuilder())
print(txt)
### ACTIONS #######################################################
k = cv2.waitKey(1) & 0xFF
if k == ord('c'):
img_temp, img, roi = initialize_images()
img = img_temp.copy()
rectangle, roi = None, None
pass
elif k == ord('r'):
# todo: ROI resizing
scale *= factor
print(scale)
img = cv2.resize(img, (0,0), fx=factor, fy=factor)
img_temp = img.copy()
elif k == 27: #esc key
cv2.destroyAllWindows()
break
def ocr(tool, img, cont='txt'):
# img is [0,1] with float64 and single channel
# langs = tool.get_available_languages()
# lang = langs[0]
if cont == 'txt':
# txt is a Python string
txt = tool.image_to_string(
Image.fromarray((img * 255.0).astype('uint8'), mode='L'),
lang="eng",
builder=ocrtools.TextBuilder()
)
return txt
if cont == 'word_boxes':
# list of box objects. For each box object:
# box.content is the word in the box
# box.position is its position on the page (in pixels)
#
# Beware that some OCR tools (Tesseract for instance)
# may return empty boxes
word_boxes = tool.image_to_string(
Image.fromarray((img * 255.0).astype('uint8'), mode='L'),
lang="eng",
builder=ocrtools.WordBoxBuilder()
)
return word_boxes
if cont == 'line_word_boxes':
# list of line objects. For each line object:
# line.word_boxes is a list of word boxes (the individual words in the line)
# line.content is the whole text of the line
# line.position is the position of the whole line on the page (in pixels)
#
# Beware that some OCR tools (Tesseract for instance)
# may return empty boxes
line_and_word_boxes = tool.image_to_string(
Image.fromarray((img * 255.0).astype('uint8'), mode='L'),
lang="eng",
builder=ocrtools.LineBoxBuilder()
)
return line_and_word_boxes
# if cont == 'digits':
# # Digits - Only Tesseract (not 'libtesseract' yet !)
# # digits is a python string
# digits = tool.image_to_string(
# Image.fromarray((img * 255.0).astype('uint8'), mode='L'),
# lang="eng",
# builder=pyocr.tesseract.DigitBuilder()
# )
# return digits
else:
raise ValueError(" Not supported OCR type ")
def ditect(self, fname):
try:
croped_img = self.__crop_img(fname)
inference_txt = self.tool.image_to_string(
croped_img,
lang='eng',
builder=b.TextBuilder(tesseract_layout=6))
except Exception as e:
sys.stderr.write(str(e))
sys.exit(0)
return inference_txt
def __init__(self):
tools = pyocr.get_available_tools()
tool = tools[0]
langs = tool.get_available_languages()
self.tool = tool
self.lang = 'jpn'
builder = builders.TextBuilder(tesseract_layout=6)
self.builder = builder
def setUp(self, get_version):
get_version.return_value = (4, 0, 0)
self.builder = builders.TextBuilder()
self.image = Image.new(mode="RGB", size=(1, 1))
self.text_file = StringIO(self._get_file_content("text"))
self.stdout = MagicMock()
self.stdout.stdout.read.return_value = b"Cuneiform for Linux 1.1.0\n"
self.stdout.wait.return_value = 0
self.tmp_filename = "/tmp/cuneiform_n0qfk87otxt"
self.enter = MagicMock()
self.enter.__enter__.return_value = MagicMock()
self.enter.__enter__.return_value.configure_mock(
name=self.tmp_filename)
def test_run_tesseract(self, popen, get_version):
message = ("Tesseract Open Source OCR Engine v4.0.0 with Leptonica\n")
self.stdout.stdout.read.return_value = message.encode()
popen.return_value = self.stdout
with tesseract.temp_dir() as tmpdir:
self.image.save(os.path.join(tmpdir, "input.bmp"))
status, error = tesseract.run_tesseract(
"input.bmp",
"output",
cwd=tmpdir,
)
self.assertEqual(status, 0)
self.assertEqual(error.decode(), message)
popen.assert_called_once_with(["tesseract", "input.bmp", "output"],
cwd=tmpdir,
startupinfo=None,
creationflags=0,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT)
get_version.return_value = (4, 0, 0)
builder = builders.TextBuilder()
with tesseract.temp_dir() as tmpdir:
self.image.save(os.path.join(tmpdir, "input2.bmp"))
status, error = tesseract.run_tesseract(
"input2.bmp",
"output2",
cwd=tmpdir,
lang="fra",
flags=builder.tesseract_flags,
configs=builder.tesseract_configs,
)
self.assertEqual(status, 0)
self.assertEqual(error.decode(), message)
popen.assert_called_with(
["tesseract", "input2.bmp", "output2", "-l", "fra", "--psm", "3"],
cwd=tmpdir,
startupinfo=None,
creationflags=0,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT)
self.assertEqual(popen.call_count, 2)
def test_init_cuneiform_params(self, get_version):
get_version.return_value = (4, 0, 0)
# XXX Maybe overkill?
# this check all combinations of parameters
for cun_dotmat, cun_fax, cun_sglcol in product(*((False, True), ) * 3):
builder = builders.TextBuilder(cuneiform_dotmatrix=cun_dotmat,
cuneiform_fax=cun_fax,
cuneiform_singlecolumn=cun_sglcol)
if cun_dotmat:
self.assertIn("--dotmatrix", builder.cuneiform_args)
else:
self.assertNotIn("--dotmatrix", builder.cuneiform_args)
if cun_fax:
self.assertIn("--fax", builder.cuneiform_args)
else:
self.assertNotIn("--fax", builder.cuneiform_args)
if cun_sglcol:
self.assertIn("--singlecolumn", builder.cuneiform_args)
else:
self.assertNotIn("--singlecolumn", builder.cuneiform_args)
def set_builder(self):
self._builder = builders.TextBuilder()
def setUp(self, get_version):
get_version.return_value = (4, 0, 0)
self.image = Image.new(mode="RGB", size=(1, 1))
self.builder = builders.TextBuilder()
def setUp(self, get_version):
get_version.return_value = (4, 0, 0)
self.builder = builders.TextBuilder()