Detect Objects In Stream - Detect Objects In Stream

在指定的影像上執行物件偵測。 支援兩個輸入法 -- (1) 上傳影像或指定影像 URL 的 (2)。 成功的回應將會以 JSON 傳回。 如果要求失敗,回應將會包含錯誤碼和訊息,以協助了解發生錯誤的原因。

POST {Endpoint}/vision/v3.2/detect
POST {Endpoint}/vision/v3.2/detect?overload=stream&model-version={model-version}

URI 參數

名稱 位於 必要 類型 Description
Endpoint
path True

string

支援的認知服務端點。

model-version
query

string

pattern: ^(latest|\d{4}-\d{2}-\d{2})(-preview)?$

選擇性參數來指定 AI 模型的版本。 接受的值包括:“latest”、“2021-04-01”、“2021-05-01”。 預設為 「latest」。。

要求標頭

Media Types: "application/octet-stream", "multipart/form-data"

名稱 必要 類型 Description
Ocp-Apim-Subscription-Key True

string

要求本文

Media Types: "application/octet-stream", "multipart/form-data"

名稱 類型 Description
Image

object (file)

影像數據流。

回應

名稱 類型 Description
200 OK

DetectResult

回應包含 JSON 格式的偵測到物件。

Other Status Codes

ComputerVisionErrorResponse

錯誤回應。

安全性

Ocp-Apim-Subscription-Key

類型: apiKey
位於: header

範例

Successful DetectObjects request

範例要求

POST https://westus.api.cognitive.microsoft.com/vision/v3.2/detect?overload=stream


"{binary}"

範例回覆

{
  "objects": [
    {
      "rectangle": {
        "x": 0,
        "y": 0,
        "w": 50,
        "h": 50
      },
      "object": "tree",
      "confidence": 0.9,
      "parent": {
        "object": "plant",
        "confidence": 0.95
      }
    }
  ],
  "requestId": "1ad0e45e-b7b4-4be3-8042-53be96103337",
  "metadata": {
    "width": 100,
    "height": 100,
    "format": "Jpeg"
  },
  "modelVersion": "2021-04-01"
}

定義

名稱 Description
BoundingRect

影像內區域的周框方塊。

ComputerVisionError

API 要求錯誤。

ComputerVisionErrorCodes

錯誤碼。

ComputerVisionErrorResponse

API 錯誤回應。

ComputerVisionInnerError

API 要求錯誤的詳細數據。

ComputerVisionInnerErrorCodeValue

錯誤碼。

DetectedObject

影像中偵測到的物件。

DetectResult

DetectImage 呼叫的結果。

ImageMetadata

影像元數據。

ObjectHierarchy

在影像內偵測到的物件。

BoundingRect

影像內區域的周框方塊。

名稱 類型 Description
h

integer (int32)

從區域左上方點測量的高度,以像素為單位。

w

integer (int32)

以像素為單位,從區域左上方點測量的寬度。

x

integer (int32)

區域左上方點的 X 座標,以像素為單位。

y

integer (int32)

區域左上方點的 Y 座標,以像素為單位。

ComputerVisionError

API 要求錯誤。

名稱 類型 Description
code

ComputerVisionErrorCodes

錯誤碼。

innererror

ComputerVisionInnerError

內部錯誤包含更具體的資訊。

message

string

說明服務所回報錯誤的訊息。

ComputerVisionErrorCodes

錯誤碼。

Description
InvalidRequest
InvalidArgument
InternalServerError
ServiceUnavailable

ComputerVisionErrorResponse

API 錯誤回應。

名稱 類型 Description
error

ComputerVisionError

錯誤內容。

ComputerVisionInnerError

API 要求錯誤的詳細數據。

名稱 類型 Description
code

ComputerVisionInnerErrorCodeValue

錯誤碼。

message

string

錯誤資訊。

ComputerVisionInnerErrorCodeValue

錯誤碼。

Description
InvalidImageFormat
UnsupportedMediaType
InvalidImageUrl
NotSupportedFeature
NotSupportedImage
Timeout
InternalServerError
InvalidImageSize
BadArgument
DetectFaceError
NotSupportedLanguage
InvalidThumbnailSize
InvalidDetails
InvalidModel
CancelledRequest
NotSupportedVisualFeature
FailedToProcess
Unspecified
StorageException

DetectedObject

影像中偵測到的物件。

名稱 類型 Description
confidence

number (double)

在影像中觀察到物件的信賴分數,其值為0到1。

object

string

對象的標籤。

parent

ObjectHierarchy

父物件,從分類法的觀點。 父對像是這個物件的較泛型形式。 例如,「鬥牛犬」會有「狗」的父代。

rectangle

BoundingRect

偵測到物件的近似位置。

DetectResult

DetectImage 呼叫的結果。

名稱 類型 Description
metadata

ImageMetadata

影像元數據。

modelVersion

string

pattern: ^(latest|\d{4}-\d{2}-\d{2})(-preview)?$

AI 模型的版本。

objects

DetectedObject[]

偵測到物件的陣列。

requestId

string

REST API 要求的標識碼。

ImageMetadata

影像元數據。

名稱 類型 Description
format

string

影像格式。

height

integer (int32)

影像高度,以像素為單位。

width

integer (int32)

影像寬度,以像素為單位。

ObjectHierarchy

在影像內偵測到的物件。

名稱 類型 Description
confidence

number (double)

在影像中觀察到物件的信賴分數,其值為0到1。

object

string

對象的標籤。

parent

ObjectHierarchy

父物件,從分類法的觀點。 父對像是這個物件的較泛型形式。 例如,「鬥牛犬」會有「狗」的父代。