365 lines
12 KiB
Python
365 lines
12 KiB
Python
|
import json
|
|||
|
|
import os
|
|||
|
|
import re
|
|||
|
|
|
|||
|
|
import cv2
|
|||
|
|
|
|||
|
|
|
|||
|
|
def split_ocr_results(ocr_results, threshold):
|
|||
|
|
"""
|
|||
|
|
对OCR识别结果进行切分，当相邻数字字符的距离大于阈值时，将文字识别框切分为两个。
|
|||
|
|
:param ocr_results: OCR识别结果，格式为文字识别框构成的List。
|
|||
|
|
:param threshold: 距离阈值，用于判断是否切分。
|
|||
|
|
:return: 切分后的OCR识别结果。
|
|||
|
|
"""
|
|||
|
|
def is_digit_pair(char1, char2):
|
|||
|
|
"""判断两个字符是否都是数字字符。"""
|
|||
|
|
return char1.isdigit() and char2.isdigit()
|
|||
|
|
|
|||
|
|
def calculate_distance(coord1, coord2):
|
|||
|
|
"""计算两个坐标之间的欧氏距离。"""
|
|||
|
|
x1, y1 = coord1
|
|||
|
|
x2, y2 = coord2
|
|||
|
|
return ((x2 - x1) ** 2 + (y2 - y1) ** 2) ** 0.5
|
|||
|
|
|
|||
|
|
def create_new_box(box_coords, char_coords, text, confidence, chars, split_index):
|
|||
|
|
"""
|
|||
|
|
创建一个新的文字识别框。
|
|||
|
|
:param box_coords: 原始文字识别框的坐标。
|
|||
|
|
:param char_coords: 单字坐标列表。
|
|||
|
|
:param text: 原始文本。
|
|||
|
|
:param confidence: 置信度。
|
|||
|
|
:param chars: 单字字符列表。
|
|||
|
|
:param split_index: 切分位置的索引。
|
|||
|
|
:return: 新的文字识别框。
|
|||
|
|
"""
|
|||
|
|
new_box_coords = [
|
|||
|
|
[box_coords[0][0], box_coords[0][1]],
|
|||
|
|
[char_coords[split_index][1][0], box_coords[0][1]],
|
|||
|
|
[char_coords[split_index][2][0], box_coords[2][1]],
|
|||
|
|
[box_coords[3][0], box_coords[3][1]]
|
|||
|
|
]
|
|||
|
|
return [
|
|||
|
|
new_box_coords,
|
|||
|
|
text[:split_index + 1],
|
|||
|
|
confidence,
|
|||
|
|
char_coords[:split_index + 1],
|
|||
|
|
chars[:split_index + 1]
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
def update_existing_box(box_coords, char_coords, text, confidence, chars, split_index):
|
|||
|
|
"""
|
|||
|
|
更新原始文字识别框。
|
|||
|
|
:param box_coords: 原始文字识别框的坐标。
|
|||
|
|
:param char_coords: 单字坐标列表。
|
|||
|
|
:param text: 原始文本。
|
|||
|
|
:param confidence: 置信度。
|
|||
|
|
:param chars: 单字字符列表。
|
|||
|
|
:param split_index: 切分位置的索引。
|
|||
|
|
:return: 更新后的文字识别框。
|
|||
|
|
"""
|
|||
|
|
return [
|
|||
|
|
[
|
|||
|
|
[char_coords[split_index + 1][0][0], box_coords[0][1]],
|
|||
|
|
[box_coords[1][0], box_coords[1][1]],
|
|||
|
|
[box_coords[2][0], box_coords[2][1]],
|
|||
|
|
[char_coords[split_index + 1][3][0], box_coords[3][1]]
|
|||
|
|
],
|
|||
|
|
text[split_index + 1:],
|
|||
|
|
confidence,
|
|||
|
|
char_coords[split_index + 1:],
|
|||
|
|
chars[split_index + 1:]
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
# 遍历每个文字识别框
|
|||
|
|
i = 0
|
|||
|
|
while i < len(ocr_results):
|
|||
|
|
ocr_box = ocr_results[i]
|
|||
|
|
box_coords, text, confidence, char_coords_list, chars = ocr_box
|
|||
|
|
# print(f"正在处理文字识别框：{text}")
|
|||
|
|
|
|||
|
|
# 遍历每个单字
|
|||
|
|
for j in range(len(chars) - 1):
|
|||
|
|
if is_digit_pair(chars[j], chars[j + 1]):
|
|||
|
|
distance = calculate_distance(char_coords_list[j][1], char_coords_list[j + 1][0])
|
|||
|
|
# print(f"字符 '{chars[j]}' 和 '{chars[j + 1]}' 之间的距离为：{distance:.2f}")
|
|||
|
|
|
|||
|
|
if distance > threshold:
|
|||
|
|
# print(f"距离大于阈值 {threshold}，进行切分。")
|
|||
|
|
# 创建新的文字识别框
|
|||
|
|
new_box = create_new_box(box_coords, char_coords_list, text, confidence, chars, j)
|
|||
|
|
# 更新原始文字识别框
|
|||
|
|
ocr_results[i] = update_existing_box(box_coords, char_coords_list, text, confidence, chars, j)
|
|||
|
|
# 插入新的文字识别框
|
|||
|
|
ocr_results.insert(i, new_box)
|
|||
|
|
# print(f"切分后的文字识别框：{new_box[1]} 和 {ocr_results[i+1][1]}")
|
|||
|
|
break # 切分后，重新处理新的文字识别框
|
|||
|
|
i += 1
|
|||
|
|
|
|||
|
|
return ocr_results
|
|||
|
|
|
|||
|
|
# 根据输入
|
|||
|
|
|
|||
|
|
|
|||
|
|
# 在ocr结果中搜索字符串
|
|||
|
|
def search_in_ocr_results(ocr_results, search_text):
|
|||
|
|
"""
|
|||
|
|
在OCR识别结果中搜索指定字符串。
|
|||
|
|
:param ocr_results: OCR识别结果。
|
|||
|
|
:param search_text: 搜索字符串。
|
|||
|
|
:return: 搜索结果。
|
|||
|
|
"""
|
|||
|
|
results = []
|
|||
|
|
for index, ocr_result in enumerate(ocr_results):
|
|||
|
|
for char_index, char in enumerate(ocr_result[4]):
|
|||
|
|
if char == search_text:
|
|||
|
|
results.append((index, char_index))
|
|||
|
|
return results
|
|||
|
|
|
|||
|
|
# 判断ocr结果是否存在
|
|||
|
|
def is_ocr_result_exist(image_path, save_path="json_data"):
|
|||
|
|
"""
|
|||
|
|
判断OCR识别结果是否存在。
|
|||
|
|
:param image_path: 图片路径。
|
|||
|
|
:return: 是否存在。
|
|||
|
|
"""
|
|||
|
|
# 获取图片名
|
|||
|
|
image_name = os.path.basename(image_path)
|
|||
|
|
# 构造JSON文件路径
|
|||
|
|
json_path = os.path.join(save_path, f"{image_name}.json")
|
|||
|
|
return os.path.exists(json_path)
|
|||
|
|
|
|||
|
|
# 将ocr结果根据图片路径（获取图片名）暂存到json文件中
|
|||
|
|
def save_ocr_result(ocr_result, image_path, save_path="json_data"):
|
|||
|
|
"""
|
|||
|
|
将OCR识别结果保存到JSON文件中。
|
|||
|
|
:param ocr_result: OCR识别结果。
|
|||
|
|
:param image_path: 图片路径。
|
|||
|
|
:return: JSON文件路径。
|
|||
|
|
"""
|
|||
|
|
# 创建保存路径
|
|||
|
|
if not os.path.exists(save_path):
|
|||
|
|
os.makedirs(save_path)
|
|||
|
|
|
|||
|
|
# 获取图片名
|
|||
|
|
image_name = os.path.basename(image_path)
|
|||
|
|
# 创建JSON文件路径
|
|||
|
|
json_path = os.path.join(save_path, f"{image_name}.json")
|
|||
|
|
# 保存OCR识别结果到JSON文件
|
|||
|
|
with open(json_path, 'w', encoding='utf-8') as file:
|
|||
|
|
json.dump(ocr_result, file, ensure_ascii=False, indent=4)
|
|||
|
|
return json_path
|
|||
|
|
|
|||
|
|
# 从json文件中读取ocr结果
|
|||
|
|
def load_ocr_result(image_path, save_path="json_data"):
|
|||
|
|
"""
|
|||
|
|
从JSON文件中加载OCR识别结果。
|
|||
|
|
:param image_path: JSON文件路径。
|
|||
|
|
:return: OCR识别结果。
|
|||
|
|
"""
|
|||
|
|
# 构造JSON文件路径
|
|||
|
|
image_name = os.path.basename(image_path)
|
|||
|
|
json_path = os.path.join(save_path, f"{image_name}.json")
|
|||
|
|
# 读取JSON文件
|
|||
|
|
with open(json_path, 'r', encoding='utf-8') as file:
|
|||
|
|
ocr_result = json.load(file)
|
|||
|
|
return ocr_result
|
|||
|
|
|
|||
|
|
# 从删除ocr结果
|
|||
|
|
def remove_ocr_result(image_path, save_path="json_data"):
|
|||
|
|
"""
|
|||
|
|
删除OCR识别结果。
|
|||
|
|
:param image_path: 图片路径。
|
|||
|
|
:return: 是否删除成功。
|
|||
|
|
"""
|
|||
|
|
# 获取图片名
|
|||
|
|
image_name = os.path.basename(image_path)
|
|||
|
|
# 构造JSON文件路径
|
|||
|
|
json_path = os.path.join(save_path, f"{image_name}.json")
|
|||
|
|
# 删除JSON文件
|
|||
|
|
if os.path.exists(json_path):
|
|||
|
|
os.remove(json_path)
|
|||
|
|
return True
|
|||
|
|
return False
|
|||
|
|
|
|||
|
|
# 绘制并保存ocr结果
|
|||
|
|
def draw_ocr_results(image_path, boxes_list, save_path="ocr_mark_data"):
|
|||
|
|
"""
|
|||
|
|
绘制OCR识别结果并保存图片。
|
|||
|
|
:param image_path: 图片路径。
|
|||
|
|
:param boxes: 文字识别框列表。
|
|||
|
|
:param save_path: 保存路径。
|
|||
|
|
:return: 保存的图片路径。
|
|||
|
|
"""
|
|||
|
|
if not os.path.exists(save_path):
|
|||
|
|
os.makedirs(save_path)
|
|||
|
|
|
|||
|
|
# 读取图片
|
|||
|
|
img = cv2.imread(image_path)
|
|||
|
|
|
|||
|
|
# 绘制方框
|
|||
|
|
for index, boxes in enumerate(boxes_list):
|
|||
|
|
for box in boxes:
|
|||
|
|
cv2.rectangle(img, (box[0], box[1]), (box[2], box[3]), get_color(index), 2)
|
|||
|
|
|
|||
|
|
|
|||
|
|
# cv2.rectangle(img, (box[0], box[1]), (box[2], box[3]), (0, 0, 255), 2)
|
|||
|
|
|
|||
|
|
# 保存图片
|
|||
|
|
save_image_path = os.path.join(save_path, os.path.basename(image_path))
|
|||
|
|
cv2.imwrite(save_image_path, img)
|
|||
|
|
return save_image_path
|
|||
|
|
|
|||
|
|
# 根据索引，获取一个颜色
|
|||
|
|
def get_color(index):
|
|||
|
|
"""
|
|||
|
|
根据索引获取一个颜色。
|
|||
|
|
:param index: 索引。
|
|||
|
|
:return: 颜色。
|
|||
|
|
"""
|
|||
|
|
colors = [
|
|||
|
|
(255, 0, 0),
|
|||
|
|
(0, 255, 0),
|
|||
|
|
(0, 0, 255),
|
|||
|
|
(255, 255, 0),
|
|||
|
|
(0, 255, 255),
|
|||
|
|
(255, 0, 255),
|
|||
|
|
(128, 0, 0),
|
|||
|
|
(0, 128, 0),
|
|||
|
|
(0, 0, 128),
|
|||
|
|
(128, 128, 0),
|
|||
|
|
(0, 128, 128),
|
|||
|
|
(128, 0, 128),
|
|||
|
|
(128, 128, 128),
|
|||
|
|
(255, 128, 0),
|
|||
|
|
(255, 0, 128),
|
|||
|
|
(128, 255, 0),
|
|||
|
|
(0, 255, 128),
|
|||
|
|
(128, 0, 255),
|
|||
|
|
(0, 128, 255),
|
|||
|
|
(255, 128, 128),
|
|||
|
|
(128, 255, 128),
|
|||
|
|
(128, 128, 255),
|
|||
|
|
(255, 255, 128),
|
|||
|
|
(255, 128, 255),
|
|||
|
|
(128, 255, 255),
|
|||
|
|
(255, 255, 255)
|
|||
|
|
]
|
|||
|
|
return colors[index % len(colors)]
|
|||
|
|
|
|||
|
|
def re_search(a, b):
|
|||
|
|
"""
|
|||
|
|
使用正则表达式在字符串列表b中搜索字符串a。
|
|||
|
|
:param a:
|
|||
|
|
:param b:
|
|||
|
|
:return:
|
|||
|
|
"""
|
|||
|
|
|
|||
|
|
if len(a) <= 1:
|
|||
|
|
return []
|
|||
|
|
|
|||
|
|
results = []
|
|||
|
|
a = replace_punctuation_with_space(a) # 删除标点符号
|
|||
|
|
|
|||
|
|
# 拷贝一份
|
|||
|
|
a_tmp = a
|
|||
|
|
|
|||
|
|
a_str = re.escape(a).replace("\\ ", "\\s*") # 将目标字符串中的空格替换为 \s*
|
|||
|
|
a_str = re.compile(a_str, re.IGNORECASE) # 忽略大小写
|
|||
|
|
# print(a_str.pattern)
|
|||
|
|
# a_str = remove_punctuation(a_str) # 删除标点符号
|
|||
|
|
for index, string in enumerate(b):
|
|||
|
|
string = replace_punctuation_with_space(string)
|
|||
|
|
match = a_str.search(string)
|
|||
|
|
if match:
|
|||
|
|
results.append((index, match.start(), match.end()))
|
|||
|
|
else:
|
|||
|
|
string = replace_punctuation_with_space(string)
|
|||
|
|
|
|||
|
|
# 如果长度小于等于1，直接跳过
|
|||
|
|
if len(string) <= 1:
|
|||
|
|
continue
|
|||
|
|
|
|||
|
|
b_str = re.escape(string).replace("\\ ", "\\s*")
|
|||
|
|
b_str = re.compile(b_str, re.IGNORECASE)
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
match = b_str.search(a)
|
|||
|
|
|
|||
|
|
if match:
|
|||
|
|
start_index = match.start()
|
|||
|
|
end_index = match.end()
|
|||
|
|
|
|||
|
|
# 将a_tmp中的匹配部分替换为空格
|
|||
|
|
a_tmp = replace_matched_text(a_tmp, b_str, "")
|
|||
|
|
|
|||
|
|
# print('匹配到的字符串:', string, '匹配到的字符串在a中的位置:', start_index, end_index)
|
|||
|
|
|
|||
|
|
results.append((index, 0, len(string)))
|
|||
|
|
|
|||
|
|
# 如果a_tmp中还有字符，说明a_tmp中的字符没有在b中匹配到
|
|||
|
|
if len(a_tmp) >= 2:
|
|||
|
|
# 重新搜索一次a_tmp
|
|||
|
|
a_tmp = re.escape(a_tmp).replace("\\ ", "\\s*")
|
|||
|
|
a_tmp = re.compile(a_tmp, re.IGNORECASE)
|
|||
|
|
# print('a_tmp:', a_tmp, 'a:', a)
|
|||
|
|
for index, string in enumerate(b):
|
|||
|
|
string = replace_punctuation_with_space(string)
|
|||
|
|
match = a_tmp.search(string)
|
|||
|
|
if match:
|
|||
|
|
# print('匹配到a_tmp:', a_tmp, 'string:', string)
|
|||
|
|
results.append((index, match.start(), match.end()))
|
|||
|
|
else:
|
|||
|
|
print('过短a_tmp:', a_tmp)
|
|||
|
|
|
|||
|
|
return results
|
|||
|
|
|
|||
|
|
|
|||
|
|
# 删除字符串中的符号
|
|||
|
|
def remove_punctuation(text):
|
|||
|
|
"""
|
|||
|
|
删除字符串中的标点符号。
|
|||
|
|
:param text: 字符串。
|
|||
|
|
:return: 删除标点符号后的字符串。
|
|||
|
|
"""
|
|||
|
|
return re.sub(r'[^\w\s]', '', text)
|
|||
|
|
|
|||
|
|
# 将字符串中的标点符号替换为空格
|
|||
|
|
def replace_punctuation_with_space(text):
|
|||
|
|
"""
|
|||
|
|
将字符串中的标点符号替换为空格。
|
|||
|
|
:param text: 字符串。
|
|||
|
|
:return: 替换后的字符串。
|
|||
|
|
"""
|
|||
|
|
return re.sub(r'[^\w\s]', ' ', text)
|
|||
|
|
|
|||
|
|
|
|||
|
|
def replace_matched_text(text, pattern, placeholder=" "):
|
|||
|
|
"""
|
|||
|
|
将被正则表达式匹配到的文本替换为指定占位符（默认为空格）。
|
|||
|
|
|
|||
|
|
参数:
|
|||
|
|
text (str): 原始字符串。
|
|||
|
|
pattern (str 或 re.Pattern): 正则表达式模式。
|
|||
|
|
placeholder (str): 用于替换匹配文本的占位符，默认为空格。
|
|||
|
|
|
|||
|
|
返回:
|
|||
|
|
str: 替换后的字符串。
|
|||
|
|
"""
|
|||
|
|
# 如果传入的是字符串形式的正则表达式，则编译它
|
|||
|
|
if isinstance(pattern, str):
|
|||
|
|
pattern = re.compile(pattern)
|
|||
|
|
|
|||
|
|
# 使用 re.sub 进行替换
|
|||
|
|
return pattern.sub(placeholder, text)
|
|||
|
|
if __name__ == "__main__":
|
|||
|
|
s = "2024：11.12"
|
|||
|
|
|
|||
|
|
print(replace_punctuation_with_space(s))
|
|||
|
|
|
|||
|
|
print(replace_punctuation_with_space("2024.11.12"))
|
|||
|
|
|
|||
|
|
print(replace_punctuation_with_space("1，375，421.63"))
|
|||
|
|
|