Vertex AI Vision 是 AI 技術輔助平台,可擷取、分析及儲存影片 資料。Vertex AI Vision 可讓使用者透過簡化的使用者介面建構及部署應用程式。
您可以使用 Vertex AI Vision 建構端對端電腦圖像解決方案,方法是利用 Vertex AI Vision 與其他主要元件 (即即時影像分析、資料串流和 Vision 倉儲) 的整合功能。Vertex AI Vision API 可讓您透過低階 API 建構高階應用程式,並建立及更新結合多個個別 API 呼叫的高階工作流程。接著,您可以向 Vertex AI Vision 平台伺服器提出單一部署要求,以單位為單位執行工作流程。
使用 Vertex AI Vision 時,您可以:
擷取即時影片 資料
使用一般和自訂的視覺 AI 模型分析資料,找出洞見
將洞察資料儲存在 Vision Warehouse 中,簡化查詢和中繼資料資訊
Vertex AI Vision 工作流程
您必須完成下列步驟才能使用 Vertex AI Vision:
擷取即時資料
Vertex AI Vision 的 架構可讓您在公用雲中快速且輕鬆地串流即時影片擷取基礎架構。
分析資料
資料擷取完成後,Vertex AI Vision 的架構可讓您輕鬆存取及調度大量的一般、自訂和專門分析模型,而且這些模型的數量還會持續增加。
在 Google Cloud,我們會優先利用 AI 原則來協助客戶透過 Vertex AI Vision 安全地開發及導入解決方案。針對 Vertex AI Vision,我們致力於依據 Google 的 AI 技術原則開發公平、公正的效能。
這項工作包括在開發期間測試偏見,例如檢視不同膚色成效,以及開發可強化隱私權並限制個人身分識別的產品功能,例如模糊處理人物和臉孔。我們致力於不斷改進及改善,並會持續將最佳做法和學習到的經驗納入 Vertex AI 產品。
當 Vertex AI Vision 整合至客戶的獨特組織背景時,可能還需要考量其他負責任 AI 相關事項。我們建議客戶在導入 Vertex AI Vision 時,特別是在建構自訂模型或 AutoML 訓練模型時,採用公平性、可解釋性、隱私權和安全性的最佳做法。在整份技術文件中,我們都提供了額外的指導和資源,協助您完成這項工作。如需更多資訊,請參閱 Google 針對負責任的 AI 做法提出的建議。
[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["難以理解","hardToUnderstand","thumb-down"],["資訊或程式碼範例有誤","incorrectInformationOrSampleCode","thumb-down"],["缺少我需要的資訊/範例","missingTheInformationSamplesINeed","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間:2025-09-04 (世界標準時間)。"],[],[],null,["# Vertex AI Vision overview\n\nVertex AI Vision is an AI-powered platform to ingest, analyze and store video\ndata. Vertex AI Vision lets users build and deploy\napplications with a simplified user interface.\n\nUsing Vertex AI Vision you can build end-to-end computer image solutions by\nleveraging Vertex AI Vision's\nintegration with other major components, namely Live Video Analytics,\ndata streams, and Vision Warehouse. The Vertex AI Vision API allows you to\nbuild a high level app from low level APIs, and create and update a high\nlevel workflow that combines multiple individual API calls. You can then\nexecute your workflow as a unit by making a single deploy request to\nthe Vertex AI Vision platform server.\n\nUsing Vertex AI Vision, you can:\n\n- Ingest real-time video data\n- Analyze data for insights using general and custom vision AI models\n- Store insights in Vision Warehouse for simplified querying and metadata information\n\nVertex AI Vision workflow\n-------------------------\n\nThe steps you complete to use Vertex AI Vision are as follows:\n\n1. **Ingest real-time data**\n\n Vertex AI Vision's architecture allows you to quickly and\n conveniently stream real-time video ingestion infrastructure in a\n public Cloud.\n2. **Analyze data**\n\n After data is ingested, Vertex AI Vision's framework provides you with easy\n access and orchestration of a large and growing portfolio of *general* ,\n *custom* ,\n \\& *specialized* analysis models.\n3. **Store and query output**\n\n After your app analyzes your data you can send this information to a\n storage destination (Vision Warehouse or BigQuery), or\n receive the data live. With Vision Warehouse you can send your app\n output to a warehouse that generalizes your search work and serves\n multiple data types and use cases.\n\n*A graph for a Vertex AI Vision occupancy analytics app in the Google Cloud console*\n\nA note on Responsible AI\n------------------------\n\nAt Google Cloud, we prioritize helping customers safely develop and implement\nsolutions using Vertex AI Vision. For Vertex AI Vision, we've worked to\ndevelop fair and equitable performance in accordance with\n[Google's AI Principles](https://ai.google/principles/).\n\nThis work includes testing for bias during development, for example looking at\nperformance across different skin tones, and developing product features to\nenhance privacy and limit personal identification, like person and face blur.\nWe are committed to iterating and improving, and we will continue to\nincorporate best practices and lessons learned into our Vertex AI\nproducts.\n\nWhen Vertex AI Vision is integrated into a customer's unique organizational\ncontext, there are likely to be additional responsible AI considerations.\nWe encourage customers to leverage fairness, interpretability, privacy and\nsecurity [best practices](https://ai.google/responsibilities/responsible-ai-practices/?category=general) when implementing Vertex AI Vision,\nespecially when building custom or AutoML trained models. Throughout this\ntechnical documentation, we have provided additional guidance and resources to\nsupport this work. To learn more, read about Google's recommendations\nfor [Responsible AI practices](https://ai.google/responsibilities/responsible-ai-practices/?category=general).\n\nWhat's next\n-----------\n\n- Read more in the blog post [\"Vertex AI Vision: Easily build and deploy computer vision\n applications at scale\"](https://cloud.google.com/blog/products/ai-machine-learning/computer-vision-for-vertex-ai).\n- Learn details about specific models in the [Occupancy analytics guide](/vision-ai/docs/occupancy-analytics-model), [Person blur guide](/vision-ai/docs/person-blur-model), [Person/vehicle detector guide](/vision-ai/docs/person-vehicle-model), or [Motion filtering guide](/vision-ai/docs/motion-filtering-model).\n- Try Vertex AI Vision in the Google Cloud console by reading the [Build an app in the console](/vision-ai/docs/build-app-console-quickstart) quickstart.\n- [Set up your local environment](/vision-ai/docs/cloud-environment) to use Vertex AI Vision."]]