直接上代码
# -*- coding: utf-8 -*- # @Author : xcl # @FILE : 1.py # @Time : 2025/2/15 7:26 # @Software : PyCharm # example.py import time import ddddocr import ddddocr from PIL import Image, ImageDraw import ddddocr file_name = '6.jpg' from io import BytesIO # 必须添加此导入 def process_image(poses): # 初始化OCR ocr = ddddocr.DdddOcr(show_ad=False) # 直接打开本地图片 1.jpg img = Image.open(file_name) # 确保图片在当前工作目录，或使用绝对路径如 "C:/path/to/1.jpg" draw = ImageDraw.Draw(img) click_identify_result = {} for row in poses: # 扩展矩形区域（防止越界） x1 = max(0, row[0] - 3) y1 = max(0, row[1] - 3) x2 = min(img.width, row[2] + 3) y2 = min(img.height, row[3] + 3) # 绘制红色边框 draw.rectangle([x1, y1, x2, y2], outline="red", width=1) # 裁剪区域 corp = img.crop((x1, y1, x2, y2)) # 保存到内存并识别 img_byte = BytesIO() corp.save(img_byte, format='PNG') img_byte.seek(0) # 重置指针 # OCR识别 word = ocr.classification(img_byte.getvalue()) click_identify_result[word] = (x1, y1, x2, y2) # img.show() # 显示结果图 return click_identify_result import cv2 det = ddddocr.DdddOcr(det=True,show_ad=False) with open(file_name, 'rb') as f: image = f.read() poses = det.detection(image) print(poses) # time.sleep(222) # poses = [(194, 31, 228, 66), (308, 106, 342, 140)] # 替换为实际坐标列表 result = process_image(poses) print("识别结果:", result) # time.sleep(2)
