AI Service > OCR > Vehicle Plate OCR > API Guide

License Plate Analysis API

Request

  • You can check the {appKey} and {secretKey} in the URL & Appkey menu at the top of the console.

[URI]

Method URI
POST https://ocr.api.nhncloudservice.com/v1.0/appkeys/{appKey}/vehicle-plate

[Request Header]

Name Value Description
Authorization {secretKey} Security key issued from the console

[Request Body]

  • Put binary data of the image file.
curl -X POST 'https://ocr.api.nhncloudservice.com/v1.0/appkeys/{appKey}/vehicle-plate' \
-F 'image=@sample.png' \
-H 'Authorization: ${secretKey}'

[Field]

Name Type Description
image multipart/form-data Image file

Response

[Response Body]

{
    "header": {
        "isSuccessful": true,
        "resultCode": 0,
        "resultMessage": "SUCCESS"
    },
    "result": {
        "fileType": "png",
        "values": [
            {
                "value":"123가4567",
                "conf":0.93
            }
        ],
        "boxes": [
            {
                "x1": 340,
                "y1": 3231,
                "x2": 523,
                "y2": 3231,
                "x3": 523,
                "y3": 3297,
                "x4": 340,
                "y4": 3297
            }
        ],
        "resolution": "normal"
    }
}

[Header]

Name Type Description
isSuccessful Boolean Analysis API success or not
resultCode Integer Result code
resultMessage String Result message (success on success, error content on failure)

[Field]

Name Type Description
fileType String File extension (jpg, png)
values List List of recognition results
values[0].value String Recognized content
values[0].conf Double Confidence score of the recognition result
resolution String normal: the resolution is the recommended resolution (HD 1280*720px) or above, low: the resolution is below the recommended resolution
boxes List List of bounding box coordinates
boxes[0] Object Coordinates of recognized area { x1, y1, x2, y2, x3, y3, x4, y4 }
  • boxes[0] Bounding box
TOP