From 54671891834809f1df77673071ff107d0652742f Mon Sep 17 00:00:00 2001 From: EvilCalf Date: Mon, 9 Dec 2019 09:50:53 +0800 Subject: [PATCH] modify some structure --- main.py | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/main.py b/main.py index 8756275..bdb3cf4 100644 --- a/main.py +++ b/main.py @@ -9,14 +9,12 @@ import os import pandas as pd -tesseract_cmd = r'.\tesseract-ocr\tesseract.exe' +tesseract_cmd = r".\tesseract-ocr\tesseract.exe" for root, dirs, files in os.walk("labimage/"): for file in files: image = Image.open(root + "/" + file) - content = pytesseract.image_to_data( - image, lang="chi_sim43", output_type="dict" - ) + content = pytesseract.image_to_data(image, lang="chi_sim43", output_type="dict") for i in range(len(content["text"])): if 0 < len(content["text"][i]): if content["text"][i] == "姓名" or ( @@ -39,17 +37,14 @@ for root, dirs, files in os.walk("labimage/"): img = cv2.imread(filename) content = pytesseract.image_to_string( img, lang="chi_sim43", output_type="dict" - ) + ) cnt = content["text"] cnt = cnt.replace(" ", "") cnt = cnt[3:] print(cnt) image.paste((0, 0, 0), (x - 10, y - 10, x + w + 400, y + h + 30)) image.save("./Output/" + file) - data = pd.DataFrame({'name': [cnt], 'dir': [root + "/" + file]}) - data.to_csv("name2file.csv",mode='a',header=False) - + data = pd.DataFrame({"name": [cnt], "dir": [root + "/" + file]}) + data.to_csv("name2file.csv", mode="a", header=False) break - - -- 2.22.0