FORMAT_CONVERSION_MAXCOMPUTE/ocr/my_infer.py

import cv2
from PIL import Image
from paddleocr import PaddleOCR
from tools.infer.utility import draw_ocr
import numpy as np
from format_convert.convert import remove_red_seal, remove_underline

# path = "../temp/complex/710.png"
# path = "../test_files/开标记录表3_page_0.png"
# path = "D:\\Project\\format_conversion\\appendix_test\\temp\\00e959a0bc9011ebaf5a00163e0ae709" + \
#         "\\00e95f7cbc9011ebaf5a00163e0ae709_pdf_page0.png"
# path = "../去章文字.jpg"
# path = "../1.jpg"
# path = "../real1.png"
path = "../temp/f1fe9c4ac8e511eb81d700163e0857b6/f1fea1e0c8e511eb81d700163e0857b6.png"
path = "../翻转1.jpg"
# 去掉公章
# image_np = cv2.imread(path)
# cv2.imshow("origin image", image_np)
# cv2.waitKey(0)
# image_np = remove_red_seal(image_np)
# cv2.imwrite("../去章文字.jpg", image_np)

# 去掉下划线
# image_np = cv2.imread(path)
# remove_underline(image_np)

with open(path, "rb") as f:
    image = f.read()
ocr_model = PaddleOCR(use_angle_cls=True, lang="ch")
image = cv2.imdecode(np.frombuffer(image, np.uint8), cv2.IMREAD_COLOR)
# # 将bgr转为rbg
np_images = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
# np_images = [cv2.imread(img_data)]
results = ocr_model.ocr(np_images, det=True, rec=True, cls=True)

bbox_list = []
text_list = []
score_list = []
for line in results:
    text_list.append(line[-1][0])
    bbox_list.append(line[0])
    score_list.append(line[-1][1])
    # print("len(text_list)", len(text_list))
    # print("len(bbox_list)", len(bbox_list))
    # print("score_list", score_list)

image = Image.fromarray(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
boxes = bbox_list

image = draw_ocr(image, boxes, text_list, score_list, drop_score=0.2)
print(type(image))
image = Image.fromarray(image)
image.show("image")