实现百度ocr功能

机器狗后台服务/.env.dev

@@ -1,3 +1,4 @@
+#--------------------------数据库配置-----------------------------------
 DB_HOST="10.0.0.17"
 DB_PORT= 3306
 DB_USER="root"
@@ -7,4 +8,11 @@ DB_CHARSET= "utf8mb4"
 DB_POOL_SIZE= 6
 DB_MAX_OVERFLOW= 10
 DB_POOL_TIMEOUT= 30
-DB_POOL_RECYCLE= 1800
+DB_POOL_RECYCLE= 1800
+#--------------------------数据库配置end-----------------------------------
+
+
+#---------------------------ocr配置----------------------------------------
+TEXT_DETECTION_MODEL_DIR= '/home/admin-root/haotian/康达瑞贝斯机器狗/PaddleOCR-3.1.0/output/PP-OCRv5_server_det_infer_20250814'
+TEXT_RECONGNITION_MODEL_DIR= '/home/admin-root/haotian/康达瑞贝斯机器狗/PaddleOCR-3.1.0/output/PP-OCRv5_server_rec_infer_20250815'
+#---------------------------ocr配置end----------------------------------------

机器狗后台服务/app/api/v1/api.py

@@ -5,9 +5,12 @@ from app.core.database import get_db
 from app.services.imageServices import ImageService
 from app.schemas.image import ImageBase
 from app.util.responseHttp import ResponseUtil
+from app.util.baiduOCR import BaiduOCR
 # from app.crud.event import event
 # from app.schemas.event import EventList, EventDetail, EventUpdate, EventQuery, TestEvent
 
+baiduOCR = BaiduOCR()
+
 router = APIRouter()
 
 @router.get("/hello")
@@ -25,3 +28,20 @@ async def test_select(
     """
     result = await ImageService.get_image_list(db,image_query)
     return ResponseUtil.success(msg="success", data=result)
+
+
+@router.get("/test_ocr")
+async def test_ocr(
+    # image_path: str = None
+):
+    """
+    测试OCR
+    """
+    
+    image_path = '/home/admin-root/haotian/康达瑞贝斯机器狗/data_image/001读表图片/2c7cc83019e7388a7041101da92c9829_frame_000000.jpg'
+    
+
+    result = baiduOCR.ocr(image_path)
+    # print(result)
+    return ResponseUtil.success(msg="success", data=result)
+    # return ResponseUtil.success(msg="success")

机器狗后台服务/app/config/config.py

@@ -20,6 +20,9 @@ class DataBaseSettings(BaseSettings):
     # class Config:
     #     env_file = ".env"
     
+class OCRSettings(BaseException):
+    TEXT_DETECTION_MODEL_DIR: str = '/home/admin-root/haotian/康达瑞贝斯机器狗/PaddleOCR-3.1.0/output/PP-OCRv5_server_det_infer_20250814'
+    TEXT_RECONGNITION_MODEL_DIR: str = '/home/admin-root/haotian/康达瑞贝斯机器狗/PaddleOCR-3.1.0/output/PP-OCRv5_server_rec_infer_20250815'
     
 class GetSettings:
     
@@ -29,6 +32,9 @@ class GetSettings:
     def get_database_settings(self):
         return DataBaseSettings()
     
+    def get_ocr_settings(self):
+        return OCRSettings()
+    
     
     
     def parse_cli_args(self):
@@ -63,4 +69,6 @@ get_settings = GetSettings()
 
 database_settings = get_settings.get_database_settings()
 
+ocr_settings = get_settings.get_ocr_settings()
+
 

机器狗后台服务/app/core/database.py

@@ -1,7 +1,7 @@
 from sqlalchemy.ext.asyncio import create_async_engine, AsyncSession  
 
 from sqlalchemy.orm import sessionmaker, declarative_base 
-from app.core.config import database_settings 
+from app.config.config import database_settings 
 
 engine = create_async_engine(
     f"mysql+aiomysql://{database_settings.DB_USER}:{database_settings.DB_PASSWORD}@{database_settings.DB_HOST}:{database_settings.DB_PORT}/{database_settings.DB_NAME}",

机器狗后台服务/app/util/baiduOCR.py

@@ -0,0 +1,32 @@
+from paddleocr import PaddleOCR
+from app.config.config import OCRSettings
+
+
+class BaiduOCR:
+    
+    def __init__(self):
+        self.model = PaddleOCR(
+        # 文本检测模型地址
+        # text_detection_model_dir = "/home/admin-root/haotian/康达瑞贝斯机器狗/ocr_model/PP-OCRv5_server_det",
+        text_detection_model_dir=OCRSettings.TEXT_DETECTION_MODEL_DIR,
+        # 文本识别模型地址
+        # text_recognition_model_dir = "/home/admin-root/haotian/康达瑞贝斯机器狗/ocr_model/PP-OCRv5_server_rec",
+        text_recognition_model_dir=OCRSettings.TEXT_RECONGNITION_MODEL_DIR,
+        use_doc_orientation_classify=False, 
+        use_doc_unwarping=False, 
+        use_textline_orientation=False)
+    def ocr(self, image_path:str):
+        result = self.model.predict(image_path)
+        result = self.parse_result(result)
+        return result 
+    
+    def parse_result(self, result):
+        text_list = []
+        for item in result:
+            text_list.append([item['rec_texts'], item['rec_scores']])
+        return text_list
+    
+    
+if __name__ == '__main__':
+    ocr = BaiduOCR()
+    print(ocr.ocr(""))