作者PHONm (USA~USA)
看板Python
標題[問題] 使用pytesseract 做ocr
時間Mon Aug 8 17:32:48 2016
我想要做字元辨識，但是字元圖像有些破裂，有些字元會變成亂碼，
所以就用OpenCV先進行一些前處理，然後存成新檔後再進行一次OCR，
但是會有UnicodeDecodeError，可是程式碼都沒有用到中文啊@@!
不曉得是否是OpenCV轉檔那邊出問題，
=====================Result=====================
<PIL.PngImagePlugin.PngImageFile image mode=RGBA size=397x112 at 0x3C0DF28>
24-D 1813f-ml 1-1
154?Dbb
<PIL.BmpImagePlugin.BmpImageFile image mode=RGB size=397x112 at 0x1131080>
Traceback (most recent call last):
File "C:/Users/cash.chien/PycharmProjects/OCR/OCRv1.1.py", line 19, in
<module>
str2 = image_to_string(img2)
File "C:\Anaconda3\lib\site-packages\pytesseract\pytesseract.py", line 167,
in image_to_string
return f.read().strip()
UnicodeDecodeError: 'cp950' codec can't decode byte 0xe2 in position 11:
illegal multibyte sequence
========================以下為原始碼=======================
from pytesseract import image_to_string
from PIL import Image
import time
import cv2
import numpy as np
img = Image.open('12.bmp')
print(img)
str = image_to_string(img)
print(str)
img1 = cv2.imread('12.bmp',1)
kernel = np.ones((3,3))
opening = cv2.morphologyEx(img1, cv2.MORPH_OPEN, kernel)
cv2.imwrite('opening.bmp',opening)
img2 = Image.open('opening.bmp')
print(img2)
str2 = image_to_string(img2)
print(str2)
感謝!
--
※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 59.124.131.189
※ 文章網址: https://webptt.com/m.aspx?n=bbs/Python/M.1470648770.A.C41.html
1F：推 Sunal: 應該是cmd的輸出問題 改成utf8試試 08/09 02:16
2F：推 goldflower: 先轉str的編碼 不過你直接把str命名掉不太好吧XD 08/11 00:22