number.png

# pip install pytesseract
# sudo apt-get install tesseract-ocr
# python
>>> import pytesseract
>>> from PIL import Image
>>> img = Image.open("/home/wq/Downloads/number.png")
>>> pytesseract.image_to_string(img)
u'1234567890'

处理机动车牌照上的数字

阈值化

>>> img.save('/home/wq/Downloads/94111.png')
>>> gray = img.convert('L')
>>> gray.save('/home/wq/Downloads/94111_gray.png')

处理后结果

>>> bw = gray.point(lambda x: 0 if x < 100 else 255, '1')
>>> bw.save('/home/wq/Downloads/94111_black.png')

处理后结果

>>> bw = bw.point(lambda x: 255-x, '1')
>>> bw.save('/home/wq/Downloads/94111_reverse.png')

处理后结果

识别数字

>>> pytesseract.image_to_string(bw)
u'4|_||'

T_T 很显然没有很好地识别出来，需要改进