Python识别快递条形码及Tesseract-OCR使用详解

#创建：2016/01/26
#文件：BarCodeIdentification.py
#作者：moverzp
#功能：识别条形码
import sys
import cv2

DECODING_TABLE = {
  '0001101': 0, '0100111': 0, '1110010': 0,
  '0011001': 1, '0110011': 1, '1100110': 1,
  '0010011': 2, '0011011': 2, '1101100': 2,
  '0111101': 3, '0100001': 3, '1000010': 3,
  '0100011': 4, '0011101': 4, '1011100': 4,
  '0110001': 5, '0111001': 5, '1001110': 5,
  '0101111': 6, '0000101': 6, '1010000': 6,
  '0111011': 7, '0010001': 7, '1000100': 7,
  '0110111': 8, '0001001': 8, '1001000': 8,
  '0001011': 9, '0010111': 9, '1110100': 9,  
  }

EDGE_TABLE = {
  2:{2:6,3:0,4:4,5:3},
  3:{2:9,3:'33',4:'34',5:5},
  4:{2:9,3:'43',4:'44',5:5},
  5:{2:6,3:0,4:4,5:3},
  }

INDEX_IN_WIDTH = (0, 4, 8, 12, 16, 20, 24, 33, 37, 41, 45, 49, 53)
def get_bar_space_width(img):
  row = img.shape[0] *1/2
  currentPix = -1
  lastPix = -1
  pos = 0
  width = []
  for i in range(img.shape[1]):#遍历一整行
    currentPix = img[row][i]
    if currentPix != lastPix:
      if lastPix == -1:
        lastPix = currentPix
        pos = i
      else:
        width.append( i - pos )
        pos = i
        lastPix = currentPix
  return width

def divide(t, l):
  if float(t) / l < 0.357:
    return 2
  elif float(t) / l < 0.500:
    return 3
  elif float(t) / l < 0.643:
    return 4
  else:
    return 5

def cal_similar_edge(data):
  similarEdge = []
  #先判断起始符
  limit = float(data[1] + data[2] + data[3] ) / 3 * 1.5
  if data[1] >= limit or data[2] >= limit or data[3] >= limit:
    return -1#宽度提取失败
  index = 4
  while index < 54:
    #跳过分隔符区间
    if index==28 or index==29 or index==30 or index==31 or index==32:
      index +=1
      continue
    #字符检测
    T1 = data[index] + data[index+1]
    T2 = data[index+1] + data[index+2]
    L = data[index] + data[index+1] + data[index+2] + data[index+3]
    similarEdge.append( divide(T1, L) )
    similarEdge.append( divide(T2, L) )
    index += 4

  return similarEdge

def decode_similar_edge(edge):
  barCode = [6]#第一个字符一定是6，中国区
  for i in range (0, 24, 2):#每个字符两个相似边，共12个字符
    barCode.append( EDGE_TABLE[edge[i]][edge[i+1]] )
  return barCode

def decode_sharp(barCode, barSpaceWidth):
  for i in range(0, 13):
    if barCode[i] == '44':
      index = INDEX_IN_WIDTH[i]
      c3 = barSpaceWidth[index+2]
      c4 = barSpaceWidth[index+3]
      if c3 > c4:
        barCode[i] = 1
      else:
        barCode[i] = 7      
    elif barCode[i] == '33':
      index = INDEX_IN_WIDTH[i]
      c1 = barSpaceWidth[index]
      c2 = barSpaceWidth[index+1]
      if c1 > c2:
        barCode[i] = 2
      else:
        barCode[i] = 8
    elif barCode[i] == '34':
      index = INDEX_IN_WIDTH[i]
      c1 = barSpaceWidth[index]
      c2 = barSpaceWidth[index+1]
      if c1 > c2:
        barCode[i] = 7
      else:
        barCode[i] = 1      
    elif barCode[i] == '43':
      index = INDEX_IN_WIDTH[i]
      c2 = barSpaceWidth[index+1]
      c3 = barSpaceWidth[index+2]
      if c2 > c3:
        barCode[i] = 2
      else:
        barCode[i] = 8

def check_bar_code(barCode):
  evens = barCode[11]+barCode[9]+barCode[7]+barCode[5]+barCode[3]+barCode[1]
  odds = barCode[10]+barCode[8]+barCode[6]+barCode[4]+barCode[2]+barCode[0]
  sum = evens * 3 + odds
  if barCode[12] == (10 - sum % 10) % 10:
    return True
  else:
    return False
#载入图像
img = cv2.imread('res\google6.jpg')
grayImg = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)#转换成单通道图像
ret, grayImg = cv2.threshold(grayImg, 200, 255, cv2.THRESH_BINARY)#二值化
grayImg = cv2.medianBlur(grayImg, 3)#中值滤波
#提取条空宽度
barSpaceWidth = get_bar_space_width(grayImg)
print 'bar & space\'s numbers:', len(barSpaceWidth)#只有60是正确的
print barSpaceWidth
#计算相似边数值
similarEdge = cal_similar_edge(barSpaceWidth)
if similarEdge == -1:
  print 'barSpaceWidth error!'
  sys.exit()
print 'similarEdge\'s numbers:', len(similarEdge)
print similarEdge
#相似边译码
barCode = decode_similar_edge(similarEdge)
#针对‘#'译码
decode_sharp(barCode, barSpaceWidth)
#校验
valid = check_bar_code(barCode)
valid = 1
print 'barcode:\n', barCode if valid else 'Check barcode error!'

height = img.shape[0]
width = img.shape[1]
cv2.line(grayImg, (0, height/2), (width, height/2),(0, 255, 0), 2)#画出扫描的行

#显示图像
cv2.imshow("origin", img)
cv2.imshow("result", grayImg)

key = cv2.waitKey(0)
if key == 27:
  cv2.destroyAllWindows()

#-*- coding:utf-8 -*-
from PIL import Image

def clean(img):
  A = img.load()
  print A
  ss = ''
  for x in xrange(img.size[0]): 
    ss += str(A[x, img.size[1]/2])
  print ss
  ls = []
  while len(ss) > 0: 
    start = ss[0]
    j = 1
    while j < len(ss) and ss[j] == start :
      j += 1
    ls.append(j)
    ss = ss[j:]
  print ls
  return ls
  #print ls
def GetUPC_A(t):
  #print t
  t = t[4:-4]
  print len(t)
  for i in xrange(len(t)):
    t[i] = (t[i] + 1) / 4
  t = t[:24] + t[29:]
  s = ''
  for i in xrange(len(t)):
   s += str(t[i]) 
  upca = ''
  for i in range(0, len(s) / 4):
    n = i * 4
    upca += dic[s[n:n + 4]]   
  print upca


dic = {'3211':'0', '2221':'1', '2122':'2', '1411':'3', '1132':'4', '1231':'5', '1114':'6', '1312':'7', '1213':'8', '3112':'9'}
img = Image.open('7.png') 
GetUPC_A(clean(img))

#-*- coding:utf-8 -*-
from PIL import Image
import pytesseract
import time

start = time.clock()#开始计时
#---------主要代码------------
im = Image.open('66.png')
code = pytesseract.image_to_string(im)
print u'验证码:' + str(code)
#---------------------------------
end = time.clock()#结束计时

print u'运行时间:' + str(end-start)

Traceback (most recent call last):
 File "111.py", line 10, in <module>
  print u'验证码:' + str(code)
UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-2: ordinal not in range(128)

import sys
reload(sys)
sys.setdefaultencoding('utf8')

# encoding=utf8
import sys
reload(sys)
sys.setdefaultencoding('utf8')