涂题卡区域检测API

使用大模型API检测图片中的涂题卡区域、下划线和表格空白单元格

GET /health

健康检查端点

响应:

{
  "status": "healthy",
  "message": "API is running"
}

POST /process_answer_sheet

处理涂题卡图片，检测涂题卡区域

请求体:

{
  "image_url": "https://example.com/image.jpg",
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 5 个涂题卡区域",
  "data": [
    {
      "boundary": [100, 200, 300, 400],
      "rows": 5,
      "columns": 4,
      "option_size": [20, 20]
    }
  ]
}

POST /process_answer_sheet_base64

处理base64编码的涂题卡图片，检测涂题卡区域

请求体:

{
  "image_base64": "base64编码的图像数据",
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 5 个涂题卡区域",
  "data": [
    {
      "boundary": [100, 200, 300, 400],
      "rows": 5,
      "columns": 4,
      "option_size": [20, 20]
    }
  ]
}

POST /find_underlines

处理base64编码的图像，检测图像中的下划线

请求体:

{
  "image_base64": "base64编码的图像数据",
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 3 条下划线",
  "data": [
    {
      "x1": 100,
      "y1": 200,
      "x2": 300,
      "y2": 250
    },
    {
      "x1": 350,
      "y1": 200,
      "x2": 550,
      "y2": 250
    },
    {
      "x1": 600,
      "y1": 200,
      "x2": 800,
      "y2": 250
    }
  ]
}

POST /find_table_cells

处理base64编码的图像，检测图像中的表格空白单元格

请求体:

{
  "image_base64": "base64编码的图像数据",
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 5 个表格空白单元格",
  "data": [
    {
      "x1": 100,
      "y1": 200,
      "x2": 300,
      "y2": 250
    },
    {
      "x1": 350,
      "y1": 200,
      "x2": 550,
      "y2": 250
    },
    {
      "x1": 600,
      "y1": 200,
      "x2": 800,
      "y2": 250
    },
    {
      "x1": 100,
      "y1": 300,
      "x2": 300,
      "y2": 350
    },
    {
      "x1": 350,
      "y1": 300,
      "x2": 550,
      "y2": 350
    }
  ]
}

POST /compare_answer_sheet

比较空白卷和答案卷，检测被选中的选项

请求体:

{
  "blank_sheet_path": "空白卷图片路径",
  "answer_sheet_path": "答案卷图片路径",
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 5 个选中的选项",
  "data": [
    {
      "row": 2,
      "column": 2
    },
    {
      "row": 3,
      "column": 4
    }
  ]
}

POST /black_pixel_analysis

比较空白卷和答案卷的黑色像素数量，检测被选中的选项

请求体:

{
  "blank_sheet_path": "空白卷图片路径",
  "answer_sheet_path": "答案卷图片路径",
  "threshold": 128,
  "increase_threshold": 30.0,
  "description": "可选的图片描述"
}

响应:

{
  "success": true,
  "message": "成功检测到 5 个选中的选项",
  "data": {
    "answer_sheet_regions": [
      {
        "boundary": [100, 200, 300, 400],
        "rows": 5,
        "columns": 4,
        "option_size": [20, 20],
        "selected_options": [
          {"row": 2, "column": 2}
        ],
        "black_pixel_counts": [
          {
            "row": 2,
            "column": 2,
            "position": [100, 200, 120, 220],
            "blank_black_pixels": 10,
            "answer_black_pixels": 50,
            "black_pixel_increase_percent": 400.0
          }
        ]
      }
    ],
    "total_regions": 1,
    "total_selected_options": 1,
    "visualization_path": "output/black_pixel_counts_comparison.png"
  }
}

使用示例

使用curl调用API (涂题卡检测):

curl -X POST "http://localhost:8000/process_answer_sheet"      -H "Content-Type: application/json"      -d '{
           "image_url": "https://marking-paper-1252908908.cos.ap-nanjing.myqcloud.com/task/20250620/60c8135fa4632f892ebe437b1750384470000001.jpg"
         }'

使用curl调用API (base64涂题卡检测):

curl -X POST "http://localhost:8000/process_answer_sheet_base64"      -H "Content-Type: application/json"      -d '{
           "image_base64": "base64编码的图像数据"
         }'

使用curl调用API (下划线检测):

curl -X POST "http://localhost:8000/find_underlines" \
     -H "Content-Type: application/json" \
     -d '{
           "image_base64": "base64编码的图像数据"
         }'

使用curl调用API (表格单元格检测):

curl -X POST "http://localhost:8000/find_table_cells" \
     -H "Content-Type: application/json" \
     -d '{
           "image_base64": "base64编码的图像数据"
         }'

使用curl调用API (答案卷比较):

curl -X POST "http://localhost:8000/compare_answer_sheet"      -H "Content-Type: application/json"      -d '{
           "blank_sheet_path": "path/to/blank_sheet.jpg",
           "answer_sheet_path": "path/to/answer_sheet.jpg"
         }'

使用curl调用API (黑色像素分析):

curl -X POST "http://localhost:8000/black_pixel_analysis"      -H "Content-Type: application/json"      -d '{
           "blank_sheet_path": "path/to/blank_sheet.jpg",
           "answer_sheet_path": "path/to/answer_sheet.jpg",
           "threshold": 128,
           "increase_threshold": 30.0
         }'

使用Python调用API (涂题卡检测):

import requests
import base64

# 方法1：使用URL
url = "http://localhost:8000/process_answer_sheet"
payload = {
    "image_url": "https://example.com/image.jpg"
}

response = requests.post(url, json=payload)
result = response.json()
print(result)

使用Python调用API (base64涂题卡检测):

import requests
import base64

# 方法2：使用base64编码
with open("image.jpg", "rb") as image_file:
    encoded_string = base64.b64encode(image_file.read()).decode('utf-8')

url = "http://localhost:8000/process_answer_sheet_base64"
payload = {
    "image_base64": encoded_string
}

response = requests.post(url, json=payload)
result = response.json()
print(result)

使用Python调用API (答案卷比较):

import requests

url = "http://localhost:8000/compare_answer_sheet"
payload = {
    "blank_sheet_path": "path/to/blank_sheet.jpg",
    "answer_sheet_path": "path/to/answer_sheet.jpg"
}

response = requests.post(url, json=payload)
result = response.json()
print(result)

使用Python调用API (黑色像素分析):

import requests

url = "http://localhost:8000/black_pixel_analysis"
payload = {
    "blank_sheet_path": "path/to/blank_sheet.jpg",
    "answer_sheet_path": "path/to/answer_sheet.jpg",
    "threshold": 128,
    "increase_threshold": 30.0
}

response = requests.post(url, json=payload)
result = response.json()
print(result)
print(f"可视化图表路径: {result['data']['visualization_path']}")

使用Python调用API (下划线检测):

import requests
import base64

# 读取图像文件并转换为base64
with open("image.jpg", "rb") as image_file:
    encoded_string = base64.b64encode(image_file.read()).decode('utf-8')

url = "http://localhost:8000/find_underlines"
payload = {
    "image_base64": encoded_string
}

response = requests.post(url, json=payload)
result = response.json()
print(result)

使用Python调用API (表格单元格检测):

import requests
import base64

# 读取图像文件并转换为base64
with open("image.jpg", "rb") as image_file:
    encoded_string = base64.b64encode(image_file.read()).decode('utf-8')

url = "http://localhost:8000/find_table_cells"
payload = {
    "image_base64": encoded_string
}

response = requests.post(url, json=payload)
result = response.json()
print(result)

查看Swagger UI文档查看ReDoc文档