OCR フォルダ振り分け 禁止文字 + 日本語 対応 【 Python 】

/* ライブラリインポート */
import os
import glob
import cv2
import sys
import configparser as cfg
import numpy as np
import pandas as pd
from operator import itemgetter
import pyocr
import pytesseract

user = os.getlogin()

/* Config設定 */
cfg = cfg.ConfigParser()
cfg.read('_config.ini', 'UTF-8')

/* OCR設定 */
tools = pyocr.get_available_tools()
if len(tools) == 0:
    print("No OCR tool found")
    sys.exit(1)
tool = tools[0]
lang = 'eng', 'jpn'
print(tool.get_name())

/* File読み込みフォルダ指定 */
files = glob.glob(cfg.get(user, 'img_folder') +
                  '/cell_cut/*.jpg', recursive=True)

/* File読み込み（複数可） */
for file in files:
    filename = os.path.basename(file)
    print('File:' + filename)
    img = cv2.imread(file, cv2.IMREAD_GRAYSCALE)
    kernel = np.ones((2, 2), np.uint8)
    img = cv2.erode(img, kernel, iterations=1)
    ret, img = cv2.threshold(img, 150, 255, cv2.THRESH_BINARY)

    minlength = int(img.shape[0] * 0.1)
    gap = 0

    lines = []
    lines = cv2.HoughLinesP(img, rho=1, theta=np.pi/360,
                            threshold=1, minLineLength=minlength, maxLineGap=gap)

    line_list = []
    k = 1
    judgelength = int(img.shape[0] * 0.95)
    for line in lines:
        x1, y1, x2, y2 = line[0]
        if abs(x1 - x2) <= k and abs(y1 - y2) >= judgelength:
            blackline = 1
            lineadd_img = cv2.line(img,
                                   (line[0][0], line[0][1]),
                                   (line[0][2], line[0][3]),
                                   (0, 0, 0),
                                   blackline)
            cv2.imwrite(cfg.get(user, 'output_folder') +
                        '/' + filename[:-4] + '_linemod.jpg', lineadd_img)
            line = (x1, y1, x2, y2)
            line_list.append(line)

    line_list.sort(key=itemgetter(0, 1, 2, 3))
    df = pd.DataFrame(line_list)

    line_count = 0
    x1 = line_list[0][0]
    base_dir = os.getcwd()
    for line in line_list:
        judge_x1 = line[0]
        judge_x2 = line[2]
        if abs(judge_x1 - x1) != 0:
            if abs(x1-judge_x1) < 2:
                x1 = judge_x1
            else:
                line_count = line_count + 1
                x2 = judge_x2
                sep_cut = img[:, x1 + 1:x2]
                w_imgh = np.ones((5, sep_cut.shape[1]), np.uint8) * 255
                sep_cut = cv2.vconcat([w_imgh, sep_cut, w_imgh])
                w_imgw = np.ones((sep_cut.shape[0], 5), np.uint8) * 255
                sep_cut = cv2.hconcat([w_imgw, sep_cut, w_imgw])

                ocr = pytesseract.image_to_string(sep_cut, lang='eng+jpn', config='--psm 10')
                ocr = ocr.replace('\n', '')
                ocr = ocr.replace('\x0c', '')
                if ocr != '':
                    print('OCR:' + str(ocr))
                    if ocr == '/' or ocr == ':' or ocr == ';' or ocr == '*' or ocr == '?' or ocr == "'" or ocr == '"' or ocr == '<' or ocr == '>' or ocr == '|' or ocr == '.' or ocr == '¥':
                        if not os.path.exists(cfg.get(user, 'char_output_folder') + cfg.get('ng_folder', ocr)):
                            os.mkdir(cfg.get(user, 'char_output_folder') + cfg.get('ng_folder', ocr))
                        os.chdir(cfg.get(user, 'char_output_folder') + cfg.get('ng_folder', ocr))
                        cv2.imwrite(filename[:-4] + '-' + str(line_count) + '_cut.jpg', sep_cut)
                        os.chdir(base_dir)
                    else:
                        if not os.path.exists(cfg.get(user, 'char_output_folder') + '/' + ocr):
                            os.mkdir(cfg.get(user, 'char_output_folder') + '/' + ocr)
                        os.chdir(cfg.get(user, 'char_output_folder') + '/' + ocr)
                        cv2.imwrite(filename[:-4] + '-' + str(line_count) + '_cut.jpg', sep_cut)
                        os.chdir(base_dir)
                x1 = judge_x1*/

/* コンフィグファイル　 */
/* config.ini */
[ng_folder]
/ = /_slash
* = /_asterisk
? = /_question
' = /_single_quotation
" = /_double_quotation
< = /_less_than_sign > = /_grater_than_sign
| = /_vertical_bar
. = /_dot
¥ = /_yen_sign