rapidocr/api

RapidOCR API

• 采用FastAPI+uvicorn实现。

使用步骤

1. 安装rapidocr_api

   $ pip install rapidocr_api
   

2. 运行
   • 用法:

     $ rapidocr_api -h
     usage: rapidocr_api [-h] [-ip IP] [-p PORT]
     
     optional arguments:
     -h, --help            show this help message and exit
     -ip IP, --ip IP       IP Address
     -p PORT, --port PORT  IP port
     

   • 示例:

     $ rapidocr_api -ip 0.0.0.0 -p 9003
     

3. 使用（本质就是发送一个POST请求，其他语言同理）
   • 以文件的方式发送请求

     import requests
     
     url = 'http://localhost:9003/ocr'
     img_path = '../python/tests/test_files/ch_en_num.jpg'
     
     with open(img_path, 'rb') as f:
         file_dict = {'image_file': (img_path, f, 'image/png')}
         response = requests.post(url, files=file_dict, timeout=60)
     
     print(response.json())
     

   • 以base64方式发送post请求

     import base64
     import requests
     
     url = 'http://localhost:9003/ocr'
     img_path = '../python/tests/test_files/ch_en_num.jpg'
     
     with open(img_path, 'rb') as fa:
         img_str = base64.b64encode(fa.read())
     
     payload = {'image_data': img_str}
     resp = requests.post(url, data=payload)
     
     print(resp.json())
     

4. API输出

   • 示例结果：

     详情

      {
          "0": {
              "rec_txt": "香港深圳抽血，",
              "dt_boxes": [
                  [265, 18],
                  [472, 231],
                  [431, 271],
                  [223, 59]
              ],
              "score": "0.8175641223788261"
          },
          "1": {
              "rec_txt": "专业查性别",
              "dt_boxes": [
                  [388, 15],
                  [636, 257],
                  [587, 307],
                  [339, 65]
              ],
              "score": "0.8293875356515249"
          },
          "2": {
              "rec_txt": "专业鉴定B超单",
              "dt_boxes": [
                  [215, 84],
                  [509, 413],
                  [453, 463],
                  [159, 134]
              ],
              "score": "0.8626169338822365"
          },
          "3": {
              "rec_txt": "b超仪器查性别",
              "dt_boxes": [
                  [128, 135],
                  [430, 478],
                  [366, 534],
                  [64, 192]
              ],
              "score": "0.8449362441897392"
          },
          "4": {
              "rec_txt": "加微信eee",
              "dt_boxes": [
                  [58, 189],
                  [268, 450],
                  [209, 498],
                  [0, 236]
              ],
              "score": "0.8176911813872201"
          },
          "5": {
              "rec_txt": "可邮寄",
              "dt_boxes": [
                  [493, 261],
                  [617, 384],
                  [577, 423],
                  [454, 300]
              ],
              "score": "0.7494261413812637"
          }
      }
     

   • 输出结果说明：

     • 如果图像中存在文字，则会输出json格式，具体格式介绍如下：

       {
        "0": {
            "rec_txt": "香港深圳抽血，",  # 识别的文本
            "dt_boxes": [  # 依次为左上角 → 右上角 → 右下角 → 左下角
                [265, 18],
                [472, 231],
                [431, 271],
                [223, 59]
            ],
            "score": "0.8175641223788261"  # 置信度
            }
       }
       

     • 如果没有检测到文字，则会输出空json({})。
5. !!说明：OCR的输出结果为最原始结果，大家可按需进一步扩展。