项目初始化

api/easy_orc_api.py

@@ -0,0 +1,61 @@
+import os
+import cv2
+import easyocr
+import numpy as np
+from core.response import SuccessResponse
+from fastapi import APIRouter, UploadFile, Form
+from schemas.orc_result import ResultInfo, ResultMain
+
+
+app = APIRouter()
+
+BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+model_dir = os.path.join(BASE_DIR, 'model')
+
+reader = easyocr.Reader(
+    ['ch_sim', 'en'],
+    model_storage_directory=model_dir,
+    download_enabled=False,
+    gpu=True)
+
+
+def preprocess_img(image):
+    # 转为灰度图
+    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+    # 二值化
+    _, binary = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
+    # 去噪声
+    denoised = cv2.medianBlur(binary, 3)
+    return denoised
+
+
+@app.post("/upload")
+async def orc(files: list[UploadFile] = Form(...)):
+    result = []
+    for file in files:
+        main = ResultMain()
+        main.file_name = file.filename
+        # 读取上传的文件内容
+        image_data = await file.read()
+
+        # 2. 将字节数据转换为 numpy 数组
+        np_array = np.frombuffer(image_data, np.uint8)
+
+        # 3. 使用 cv2.imdecode 解码为 OpenCV 图像
+        image = cv2.imdecode(np_array, cv2.IMREAD_COLOR)
+        new_image = preprocess_img(image)
+        datas = reader.readtext(np.array(new_image))
+        infos = []
+        for data in datas:
+            info = ResultInfo()
+            bounding_boxs = data[0]
+            left_up = bounding_boxs[0]
+            right_down = bounding_boxs[2]
+            info.bounding_box_left_up = [int(left_up[0]), int(left_up[1])]
+            info.bounding_box_right_down = [int(right_down[0]), int(right_down[1])]
+            info.text = data[1]
+            info.confidence = round(data[2], 4)
+            infos.append(info)
+        main.infos = infos
+        result.append(main.model_dump())
+    return SuccessResponse(data=result)

api/paddle_ocr_api.py

@@ -0,0 +1,55 @@
+from core.response import SuccessResponse
+from fastapi import APIRouter, UploadFile, Form
+from schemas.orc_result import ResultInfo, ResultMain
+
+from paddleocr import PaddleOCR
+import numpy as np
+import cv2
+
+
+app = APIRouter()
+
+
+paddle_ocr = PaddleOCR(lang='ch')
+
+
+def enhance_image(img):
+    """增强亮部和暗部，提升数字清晰度"""
+    lab = cv2.cvtColor(img, cv2.COLOR_BGR2LAB)
+    l, a, b = cv2.split(lab)
+    clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8,8))
+    cl = clahe.apply(l)
+    enhanced = cv2.merge((cl, a, b))
+    return cv2.cvtColor(enhanced, cv2.COLOR_LAB2BGR)
+
+
+@app.post("/upload")
+async def orc(files: list[UploadFile] = Form(...)):
+    result = []
+    for file in files:
+        main = ResultMain()
+        main.file_name = file.filename
+        # 读取上传的文件内容
+        image_data = await file.read()
+        # 从字节数据读取图像
+        np_array = np.frombuffer(image_data, np.uint8)
+        img = cv2.imdecode(np_array, cv2.IMREAD_COLOR)
+        # 图像增强
+        img = enhance_image(img)
+        # 4. 调用 OCR 模型进行识别
+        datas = paddle_ocr.ocr(img, cls=False)
+        infos = []
+        if datas:
+            for data in datas[0]:
+                info = ResultInfo()
+                bounding_boxs = data[0]
+                left_up = bounding_boxs[0]
+                right_down = bounding_boxs[2]
+                info.bounding_box_left_up = [int(left_up[0]), int(left_up[1])]
+                info.bounding_box_right_down = [int(right_down[0]), int(right_down[1])]
+                info.text = data[1][0]
+                info.confidence = round(data[1][1], 4)
+                infos.append(info)
+            main.infos = infos
+            result.append(main.model_dump())
+    return SuccessResponse(data=result)