OCR提取+识别方案

1. 内容提取

通过 YOLO 提取需要识别的区域

1.1 安装ultralytics

1. 创建虚拟环境(可选)

   # 创建虚拟环境
   python -m venv ultralytics-env
   
   # 激活虚拟环境
   ### 激活虚拟环境将更改 shell 的提示以显示您正在使用的虚拟环境，并修改环境，以便运行时 python可以获得特定版本和安装的 Python。例如：
   source ultralytics-env/bin/activate
   
   # 显示虚拟环境中安装的所有软件包：
   python -m pip list
   
   # 停用/退出虚拟环境
   # deactivate

2. 配置阿里云加速

   # 配置 Pip 清华镜像源，--user参数表示当前用户生效
   pip3 config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple --user
   pip3 config set install.trusted-host pypi.tuna.tsinghua.edu.cn --user
   
   # 或者aliyu源： https://mirrors.aliyun.com/pypi/simple
   # 国外可以使用官方源: https://pypi.org/simple
   
   
   # 查看 Pip 所有配置项, 确认代理配置成功
   pip3 config list --user

3. 安装ultralytics

   # Install the ultralytics package from PyPI
   pip install ultralytics
   
   # 导出依赖项
   # pip freeze > requirements.txt
   # 安装依赖项
   # pip install -r requirements.txt

   官方文档: https://docs.ultralytics.com/quickstart/

1.2 编写脚本

cut.py

from ultralytics import YOLO

# Load a model
model = YOLO("best.pt")  # pretrained YOLOv8n model

# Run batched inference on a list of images
modelDir = "Downloads/"

# Run inference on 'bus.jpeg' with arguments
results = model.model(modelDir + "bus.jpeg", save=True, imgsz=96)

# Process results list
for result in results:
    boxes = result.boxes  # Boxes object for bounding box outputs
    masks = result.masks  # Masks object for segmentation masks outputs
    keypoints = result.keypoints  # Keypoints object for pose outputs
    probs = result.probs  # Probs object for classification outputs
    obb = result.obb  # Oriented boxes object for OBB outputs
    result.show()  # display to screen
    result.save(filename="result.jpg")  # save to disk
    result.save_txt(txt_file="result.txt")
    print(result.tojson())

执行检测
python cut.py

2. OCR识别

pip install paddlepaddle-gpu