개정판 eea44d8f
issue #663: fixed a routine recognizing text from image
Change-Id: I6360bae8fe30c0863a5fb0dfa15fee64f2970af3
| DTI_PID/DTI_PID/TextDetector.py | ||
|---|---|---|
| 39 | 39 |
try: |
| 40 | 40 |
tInfoList = self.getTextAreaInfo(img, offset[0], offset[1]) |
| 41 | 41 |
except Exception as ex: |
| 42 |
print('error occured({}) in {}:{}'.format(ex, sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno))
|
|
| 42 |
print('error occurred({}) in {}:{}'.format(ex, sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno))
|
|
| 43 | 43 |
|
| 44 | 44 |
return tInfoList |
| 45 | 45 |
|
| ... | ... | |
| 202 | 202 |
2018.11.08 euisung add white char list check process on db |
| 203 | 203 |
''' |
| 204 | 204 |
@staticmethod |
| 205 |
def recognizeTextFromImage(tInfo, imgOCR, offset, searchedSymbolList, worker, listWidget, maxProgressValue): |
|
| 205 |
def recognizeTextFromImage(tInfos, imgOCR, offset, searchedSymbolList, worker, listWidget, maxProgressValue):
|
|
| 206 | 206 |
import re |
| 207 | 207 |
res = [] |
| 208 | 208 |
|
| 209 |
appDocData = AppDocData.instance()
|
|
| 209 |
app_doc_data = AppDocData.instance()
|
|
| 210 | 210 |
|
| 211 | 211 |
try: |
| 212 |
x = tInfo.getX() - round(offset[0]) |
|
| 213 |
y = tInfo.getY() - round(offset[1]) |
|
| 214 |
img = imgOCR[y:y+tInfo.getH(), x:x+tInfo.getW()] |
|
| 215 |
|
|
| 216 |
# set angle 0 if symbol contains the text area is instrumentation |
|
| 217 |
category = None |
|
| 218 |
contains = [symbol for symbol in searchedSymbolList if symbol.contains(tInfo)] |
|
| 219 |
if contains: |
|
| 220 |
_type = contains[0].getType() |
|
| 221 |
category = appDocData.getSymbolCategoryByType(_type) |
|
| 222 |
if 'Instrumentation' == category: tInfo.setAngle(0) |
|
| 223 |
# up to here |
|
| 212 |
for tInfo in tInfos: |
|
| 213 |
x = tInfo.getX() - round(offset[0]) |
|
| 214 |
y = tInfo.getY() - round(offset[1]) |
|
| 215 |
img = imgOCR[y:y+tInfo.getH(), x:x+tInfo.getW()] |
|
| 216 |
|
|
| 217 |
# set angle 0 if symbol contains the text area is instrumentation |
|
| 218 |
category = None |
|
| 219 |
contains = [symbol for symbol in searchedSymbolList if symbol.contains(tInfo)] |
|
| 220 |
if contains: |
|
| 221 |
_type = contains[0].getType() |
|
| 222 |
category = app_doc_data.getSymbolCategoryByType(_type) |
|
| 223 |
if 'Instrumentation' == category: |
|
| 224 |
tInfo.setAngle(0) |
|
| 225 |
# up to here |
|
| 224 | 226 |
|
| 225 |
whiteCharList = appDocData.getConfigs('Text Recognition', 'White Character List')
|
|
| 226 |
if len(whiteCharList) is 0: |
|
| 227 |
resultTextInfo = TOCR.getTextInfo(img, (x, y), tInfo.getAngle(), language=appDocData.OCRData)
|
|
| 228 |
else: |
|
| 229 |
resultTextInfo = TOCR.getTextInfo(img, (x, y), tInfo.getAngle(), language=appDocData.OCRData, conf = whiteCharList[0].value)
|
|
| 227 |
whiteCharList = app_doc_data.getConfigs('Text Recognition', 'White Character List')
|
|
| 228 |
if len(whiteCharList) is 0:
|
|
| 229 |
resultTextInfo = TOCR.getTextInfo(img, (x, y), tInfo.getAngle(), language=app_doc_data.OCRData)
|
|
| 230 |
else:
|
|
| 231 |
resultTextInfo = TOCR.getTextInfo(img, (x, y), tInfo.getAngle(), language=app_doc_data.OCRData, conf=whiteCharList[0].value)
|
|
| 230 | 232 |
|
| 231 |
if resultTextInfo is not None and len(resultTextInfo) > 0: |
|
| 232 |
for result in resultTextInfo: |
|
| 233 |
result.setX(result.getX() + round(offset[0])) |
|
| 234 |
result.setY(result.getY() + round(offset[1])) |
|
| 235 |
if 'Instrumentation' == category: |
|
| 236 |
text = re.sub('[^a-zA-Z0-9]+', '', result.getText())
|
|
| 237 |
result.setText(text) |
|
| 238 |
res.extend(resultTextInfo) |
|
| 239 |
|
|
| 240 |
if listWidget is not None: |
|
| 241 |
item = QListWidgetItem('{},{},{} is recognized'.format(resultTextInfo[0].getX(), resultTextInfo[0].getY(), resultTextInfo[0].getText()))
|
|
| 242 |
listWidget.addItem(item) |
|
| 243 |
else: |
|
| 244 |
pass |
|
| 233 |
if resultTextInfo is not None and len(resultTextInfo) > 0: |
|
| 234 |
for result in resultTextInfo: |
|
| 235 |
result.setX(result.getX() + round(offset[0])) |
|
| 236 |
result.setY(result.getY() + round(offset[1])) |
|
| 237 |
if 'Instrumentation' == category: |
|
| 238 |
text = re.sub('[^a-zA-Z0-9]+', '', result.getText())
|
|
| 239 |
result.setText(text) |
|
| 240 |
|
|
| 241 |
res.extend(resultTextInfo) |
|
| 245 | 242 |
|
| 246 |
if worker is not None: worker.updateProgress.emit(maxProgressValue, resultTextInfo[0].getText() if resultTextInfo is not None and 1 == len(resultTextInfo) else None) |
|
| 243 |
if listWidget is not None: |
|
| 244 |
item = QListWidgetItem('{},{},{} is recognized'.format(resultTextInfo[0].getX(), resultTextInfo[0].getY(), resultTextInfo[0].getText()))
|
|
| 245 |
listWidget.addItem(item) |
|
| 246 |
else: |
|
| 247 |
pass |
|
| 248 |
|
|
| 249 |
if worker is not None: |
|
| 250 |
worker.updateProgress.emit(maxProgressValue, resultTextInfo[0].getText() if resultTextInfo is not None and 1 == len(resultTextInfo) else None) |
|
| 247 | 251 |
except Exception as ex: |
| 248 |
message = 'error occured({}) in {}:{}'.format(ex, sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno)
|
|
| 249 |
worker.displayLog.emit(MessageType.Error, message) |
|
| 252 |
message = 'error occurred({}) in {}:{}'.format(repr(ex), sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno)
|
|
| 253 |
if worker is not None: |
|
| 254 |
worker.displayLog.emit(MessageType.Error, message) |
|
| 250 | 255 |
|
| 251 | 256 |
return res |
| 252 | 257 |
|
| ... | ... | |
| 267 | 272 |
''' |
| 268 | 273 |
def recognizeText(self, imgSrc, offset, tInfoList, searchedSymbolList, worker, listWidget, maxProgressValue, onlyTextArea = False): |
| 269 | 274 |
import concurrent.futures as futures |
| 275 |
from multiprocessing import Process, Queue |
|
| 270 | 276 |
from Area import Area |
| 271 | 277 |
|
| 272 | 278 |
try: |
| ... | ... | |
| 282 | 288 |
imgOCR = cv2.imread(path, 1) |
| 283 | 289 |
imgOCR = cv2.threshold(cv2.cvtColor(imgOCR, cv2.COLOR_BGR2GRAY), 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)[1] |
| 284 | 290 |
|
| 285 |
pool = futures.ThreadPoolExecutor(max_workers = THREAD_MAX_WORKER) |
|
| 286 |
for tInfo in tInfoList: |
|
| 287 |
future = pool.submit(TextDetector.recognizeTextFromImage, tInfo, imgOCR, offset, searchedSymbolList, worker, listWidget, maxProgressValue) |
|
| 291 |
text_info_array = np.array_split(tInfoList, THREAD_MAX_WORKER) |
|
| 292 |
pool = futures.ThreadPoolExecutor(max_workers=THREAD_MAX_WORKER) |
|
| 293 |
for tInfo in text_info_array: |
|
| 294 |
future = pool.submit(TextDetector.recognizeTextFromImage, tInfo, imgOCR, offset, searchedSymbolList, |
|
| 295 |
worker, listWidget, maxProgressValue) |
|
| 288 | 296 |
data = future.result() |
| 289 |
if data: self.textInfoList.extend(data) |
|
| 290 |
pool.shutdown(wait = True) |
|
| 297 |
if data: |
|
| 298 |
self.textInfoList.extend(data) |
|
| 299 |
pool.shutdown(wait=True) |
|
| 291 | 300 |
|
| 292 |
## remove text item has only 1 character
|
|
| 293 |
#for index in range(len(self.textInfoList) - 1, -1, -1): |
|
| 301 |
# remove text item has only 1 character |
|
| 302 |
# for index in range(len(self.textInfoList) - 1, -1, -1):
|
|
| 294 | 303 |
# if len(self.textInfoList[index].getText()) is 1: |
| 295 | 304 |
# self.textInfoList.pop(index) |
| 296 | 305 |
|
| ... | ... | |
| 331 | 340 |
|
| 332 | 341 |
if worker is not None: worker.updateProgress.emit(maxProgressValue, None) |
| 333 | 342 |
except Exception as ex: |
| 334 |
message = 'error occured({}) in {}:{}'.format(ex, sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno)
|
|
| 343 |
message = 'error occurred({}) in {}:{}'.format(repr(ex), sys.exc_info()[-1].tb_frame.f_code.co_filename, sys.exc_info()[-1].tb_lineno)
|
|
| 335 | 344 |
worker.displayLog.emit(MessageType.Error, message) |
| 336 | 345 |
|
| 337 | 346 |
''' |
내보내기 Unified diff
