[Pycodeocr-main] SF.net SVN: pycodeocr:[65] branches/redesign/Recognition
Status: Beta
Brought to you by:
drtrigon
Menu
▾
▴
[Pycodeocr-main] SF.net SVN: pycodeocr:[65] branches/redesign/Recognition
|
From: <la...@us...> - 2014-08-24 10:19:39
|
Revision: 65
http://sourceforge.net/p/pycodeocr/code/65
Author: laserb
Date: 2014-08-24 10:19:30 +0000 (Sun, 24 Aug 2014)
Log Message:
-----------
add more documentation
Modified Paths:
--------------
branches/redesign/Recognition/Barcode.py
branches/redesign/Recognition/ESR.py
branches/redesign/Recognition/ImageRecognition.py
branches/redesign/Recognition/PDF417.py
Modified: branches/redesign/Recognition/Barcode.py
===================================================================
--- branches/redesign/Recognition/Barcode.py 2014-08-24 10:18:01 UTC (rev 64)
+++ branches/redesign/Recognition/Barcode.py 2014-08-24 10:19:30 UTC (rev 65)
@@ -9,7 +9,7 @@
check = not (data['type'] == "unknown")
if check:
check = check and ( int(data['crc']) == 0 ) # CRC error check?
- check = check and (float(data['error']) < 0.15) # recognizion errors?
+ check = check and (float(data['error']) < 0.15) # recognition errors?
print 'type:' +" "+ data['type']
print 'chars:' +" "+ data['chars']
print 'crc:' +" "+ data['crc']
@@ -17,7 +17,7 @@
data = data['code']
else:
data = "type: " + data['type']
- return check
+ return (check, data)
def setImage(self,image):
self.image = image
Modified: branches/redesign/Recognition/ESR.py
===================================================================
--- branches/redesign/Recognition/ESR.py 2014-08-24 10:18:01 UTC (rev 64)
+++ branches/redesign/Recognition/ESR.py 2014-08-24 10:19:30 UTC (rev 65)
@@ -1,62 +1,104 @@
# -*- coding: utf8 -*-
+##
+# @file ESR.py
+# @package ImageRecognition
+# @authors drtrigon, laserb
+# @version 2.0b
+# @brief Recognize codeline from ESR
+#
+# $Id: $
+#
+#
+# @section Description
+#
+# Extract information from ESR.
+#
+# @section Dependencies
+#
+# @li Python modules needed are: os, re.
+# @li Modules: utils, Recognition
+#
from Recognition import ImageRecognition
import utils
import re, os
+## Extract information from ESR
class ESR(ImageRecognition.ImageRecognitionBase):
+ ## init
def __init__(self):
super(ESR,self).__init__("ESR")
self.temp_name = "image_object_temp_file"
self.valid_code_len = [ 53, 43, 42, 32 ] # valid code lenghts
-
+ ## @overwrite verification
def verify(self):
+ # correct for common char recognition errors
self.recognitionData.setContent(self.char_correction(self.recognitionData.getContent()))
+ # get data
data = self.recognitionData.getContent()
+
+ # check for invalid characters
check = (not "?" in data) or (not "!" in data) # any unrecognized char in code?
check = check and ( len(data) in self.valid_code_len ) # correct code len?
if not check:
return (False, "Invalid length or invalid characters")
+ # extract amount, reference and account from codeline
tmp = data[:-1].split(">")
amount = tmp[0]
tmp = tmp[1].split("+ ")
reference = tmp[0]
account = tmp[1]
+ # do checksum for amount, reference and account
if not self.checksum(amount):
return (False, "Checksum for amount failed.")
if not self.checksum(reference):
return (False, "Checksum for reference failed.")
if not self.checksum(account):
return (False, "Checksum for account failed.")
+
+ # return true if all checks are positive
return (True, "")
# thanks to http://www.hosang.ch/modulo10.aspx
# much more details http://www.sic.ch/de/dl_tkicch_dta.pdf ( page 51 )
# @see http://www.bundesbank.de/download/zahlungsverkehr/zv_pz201012.pdf
# @see http://www.bundesbank.de/zahlungsverkehr/zahlungsverkehr_pruefziffernberechnung.php
+ # @return True if checksum is valid, false otherwise
def checksum(self, number):
tabelle = [0,9,4,6,8,2,7,1,3,5]
uebertrag = 0
+
# iterate over each character
for i in range(len(number)-1):
uebertrag = tabelle[(uebertrag + int(number[i]) )%10]
uebertrag = (10-uebertrag)%10
+
+ # check if last digit equals checksum
return str(uebertrag) == number[-1]
+ # @overwrite recognizeImage
def recognizeImage(self, image, rotate = None):
super(ESR,self).recognizeImage(image)
+
+ # if no rotation is set, use default: [0, 90, 90, 90]
if rotate == None:
rotate = self.rotate
+
+ # try all rotation and test if data is valid
for angle in rotate:
+ # convert image
self.recognitionData.convert({"-depth":str(self.depth),"-rotate":str(angle)}, "convert: unable to open image")
+
+ # do recognition with tesseract
self.recognitionData.getImage().save(self.temp_name + ".png")
runExternal = utils.RunExternal("tesseract %s %s" % (self.temp_name + ".png",self.temp_name+"tesseract"),error_msg="Couldn't find text")
runExternal()
+
# Simple workaround: The program simply continuous if no text is found at all
# and therefore no txt-file is generated.
if not os.path.exists(self.temp_name+"tesseract.txt"):
@@ -67,10 +109,16 @@
df.close()
os.remove(self.temp_name + "tesseract.txt") # clean-up
os.remove(self.temp_name + ".png") # clean-up
+
+ # set content
self.recognitionData.setContent(data)
+
+ # check if data is valid
(check, msg) = self.verify()
if check:
break
+
+ # return the data
return self.recognitionData.getContent()
## Character correction after recogition (on basis that there should be numbers and few special chars).
Modified: branches/redesign/Recognition/ImageRecognition.py
===================================================================
--- branches/redesign/Recognition/ImageRecognition.py 2014-08-24 10:18:01 UTC (rev 64)
+++ branches/redesign/Recognition/ImageRecognition.py 2014-08-24 10:19:30 UTC (rev 65)
@@ -1,70 +1,104 @@
# -*- coding: utf8 -*-
##
-# @file File.py
-# @package Source
-# @extends SourceBase
+# @file ImageRecognition.py
+# @package ImageRecognition
# @authors drtrigon, laserb
# @version 2.0b
-# @brief Base class for sources.
+# @brief Base class for image recognition.
#
# $Id: $
#
#
# @section Description
#
-# Get an image from file.
+# Extract information from image.
#
# @section Dependencies
#
+# @li Python modules needed are: os, gtk, PIL.
+# @li Modules: utils
#
import os, gtk
from PIL import Image
import utils
+## Base class for image recognition. This class is not useful by itself.
+# Extend this class for a specific recognition technique.
class ImageRecognitionBase(object):
+ ## init
def __init__(self, name):
self.name = name
self.depth = 8
self.rotate = [0,90,90,90]
+ ## Verify if data is valid.
+ # @pre image acquired and recognized
+ # @post information extracted
+ # @return (boolean, data) True if data is valid, false otherwise
+ # if data is valid return data, otherwise return error message
def verify(self):
return (True, "")
+ ## get image and associated data
+ # @pre image recognized
+ # @return image and asscociated data
def getRecognitionData(self):
return self.recognitionData
+ ## recognize image
+ # @pre image acquired
+ # @param image
+ # @post information is extracted
+ # @return extracted information
def recognizeImage(self,image):
recognitionData = RecognitionData(image)
self.recognitionData = recognitionData
return self.recognitionData.getContent()
+ ## get name
+ # @return name
def getName(self):
return self.name
+ ## get rotate
+ # @return rotate
def getRotate(self):
return self.rotate
+ ## set rotate
+ # @param rotation angle
def setRotate(self, rotate):
self.rotate = rotate
+ ## get depth
+ # @return depth
def getDepth(self):
return self.depth
+ ## set depth
+ # @param depth
def setDepth(self, depth):
self.depth = depth
-
+## store for image and associated data
class RecognitionData(object):
+ ## init
def __init__(self, image):
self.image = image
self.temp_name = "image_object_temp_file"
self.content = ""
+ ## get image
+ # @return image
def getImage(self):
return self.image
+ ## convert initial image to prepare it for recognition
+ # @pre image acquired
+ # @param convertion options, error message
+ # @post image optimized for recognition
def convert(self,options,msg):
path = os.path.abspath(".")
orig = os.path.join(path,self.temp_name + ".png")
@@ -80,12 +114,21 @@
os.remove(orig) # clean-up
os.remove(new) # clean-up
+ ## set information content of image
+ # @pre image recognized
+ # @param information from image
+ # @post information set
def setContent(self, content):
self.content = content
+ ## get information associated with image
+ # @return information content
def getContent(self):
return self.content
+ ## save content to clipboard
+ # @pre information set
+ # @post information saved in clipboard
def saveClipboard(self):
# get the clipboard
clipboard = gtk.clipboard_get()
@@ -94,21 +137,35 @@
# make our data available to other applications
clipboard.store()
-
+## Run the recognition and verify the information
class RunRecognition(object):
+ ## init
def __init__(self, source, data, imageRecognition):
self.source = source
self.data = data
self.imageRecognition = imageRecognition
+ ## call
+ # @return (boolean, data), True if data is valid, False otherwise
+ # data if the acquired data is image, error-message otherwise
def __call__(self):
image = self.source.getImage(self.data)
return self.recognize(image)
+ ## recognize image
+ # @return (boolean, data), True if data is valid, False otherwise
+ # data if the acquired data is image, error message otherwise
def recognize(self, image):
+ # do recognition
self.imageRecognition.recognizeImage(image)
+
+ # verify data
(check, msg) = self.imageRecognition.verify()
+
+ # if data is valid, save to clipboard and return data
if check:
self.imageRecognition.getRecognitionData().saveClipboard()
return (True, self.imageRecognition.getRecognitionData().getContent())
+
+ # if data is invalid return error message
return (False, msg)
\ No newline at end of file
Modified: branches/redesign/Recognition/PDF417.py
===================================================================
--- branches/redesign/Recognition/PDF417.py 2014-08-24 10:18:01 UTC (rev 64)
+++ branches/redesign/Recognition/PDF417.py 2014-08-24 10:19:30 UTC (rev 65)
@@ -5,7 +5,7 @@
super(PDF417,self).__init("PDF417")
def verify(self):
- return True
+ return (True, "")
def setImage(self,image):
self.image = image
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
