公告：凡是在
2025 年 4 月 15 日前註冊使用 Earth Engine 的非商業專案，都必須
驗證非商業用途資格，才能繼續存取 Earth Engine。
關於 Google Earth Engine
透過集合功能整理內容
你可以依據偏好儲存及分類內容。
Google Earth Engine 是Google Cloud 產品，可用於大規模地理空間分析。這個平台結合了數 PB 規模的衛星圖像和地理空間資料集目錄，並採用行星級運算技術，加快環境研究和應用的速度。
主要功能
簡化且可擴充的地理空間分析
Earth Engine 將豐富的地理空間資料目錄與分散式運算整合，可透過用戶端程式庫存取。使用者可以存取各種衛星和環境資料，並整合自己的資料集。這個平台會根據使用者指定的參數，自動處理資料投影、縮放和合成作業，簡化地理空間分析。其分析函式可在不同規模下高效運作，無須明確的資料準備步驟或區塊。Earth Engine 會在內部管理複雜的資料處理和運算縮放作業，讓使用者能專注於分析，而非技術設定。
處理環境
Earth Engine 支援兩種分析模式：
- 互動模式：快速探索少量資料並即時製作資料視覺化圖表。
- 批次模式：適用於大量資料的大型運算密集工作。
開發環境
開發人員可以選擇兩種主要開發環境：
- Python 用戶端程式庫：Earth Engine 的彈性介面，可與更廣泛的 Python 生態系統整合，在 Jupyter 筆記本中進行互動式分析及支援進階工作流程。
- JavaScript 程式碼編輯器：專用的網路式開發環境，可用於快速建立原型、探索及建立 Earth Engine 應用程式。
視覺化和結果
Earth Engine 支援從初始原型設計到最終資料匯出的地理空間分析作業。這項工具整合了互動式地圖小工具，並提供高效率的平鋪和運算系統，可在程式碼編輯器和 Python 環境中提供快速的視覺化和檢查功能。這樣一來，您就能立即探索資料並進行疊代。準備就緒後，使用者可以匯出光柵和向量結果至 Google Cloud Storage、BigQuery 或 Google 雲端硬碟，也可以在本機下載與 pandas、NumPy 和 Xarray 相容的格式資料。此外，Earth Engine 也支援互動式網頁應用程式的建立，讓使用者能與廣大觀眾分享地理空間洞察資料。
機器學習
Earth Engine 內建用於迴歸、分類、圖像區隔和準確度評估的機器學習工具。訓練完成後，模型即可儲存並重複套用。傳統的機器學習工作流程在 Earth Engine 的整合系統中經過簡化處理。如需進階選項或外部訓練的模型，可以整合 Vertex AI，將模型帶入 Earth Engine 資料，或建立深度學習模型和以神經網路為基礎的分析。
存取權和管理
Earth Engine 可用於商業和非商業用途。非商業用途可免費使用，但商業用途須支付訂閱費和運算費用。所有運算和私人資料都會與 Google Cloud 專案建立關聯，讓使用者透過 Google Cloud 主控台控管存取權、資源管理和用量監控。這項整合功能可讓您集中管理專案、取得詳細的帳單資訊，並應用 Google Cloud 強大的安全性和法規遵循功能。使用者可以利用 Identity and Access Management (IAM) 控管權限，並透過 Cloud Monitoring 和 Cloud Logging 記錄活動及監控資源使用情形。
除非另有註明，否則本頁面中的內容是採用創用 CC 姓名標示 4.0 授權，程式碼範例則為阿帕契 2.0 授權。詳情請參閱《Google Developers 網站政策》。Java 是 Oracle 和/或其關聯企業的註冊商標。
上次更新時間：2025-02-18 (世界標準時間)。
[null,null,["上次更新時間：2025-02-18 (世界標準時間)。"],[[["\u003cp\u003eGoogle Earth Engine is a cloud-based platform that provides petabytes of satellite imagery and geospatial datasets for environmental analysis.\u003c/p\u003e\n"],["\u003cp\u003eIt offers tools for geospatial analysis, including interactive and batch processing modes, as well as Python and JavaScript development environments.\u003c/p\u003e\n"],["\u003cp\u003eUsers can visualize and export results to various formats and platforms, including Google Cloud Storage, BigQuery, and Google Drive.\u003c/p\u003e\n"],["\u003cp\u003eEarth Engine incorporates machine learning capabilities for tasks like regression, classification, and image segmentation, and integrates with Vertex AI for advanced modeling.\u003c/p\u003e\n"],["\u003cp\u003eAccess is available for both commercial and non-commercial use, with options for managing projects, resources, and permissions through Google Cloud.\u003c/p\u003e\n"]]],["Google Earth Engine enables scalable geospatial analysis by combining a vast data catalog with planetary-scale computation. Users can access, process, and analyze satellite and environmental data using Python or JavaScript. It supports both interactive and batch processing for tasks. Results can be visualized, exported to various platforms (Google Cloud Storage, BigQuery, etc.), or integrated into interactive web applications. Machine learning tools are included, and Vertex AI integration is available for advanced models. Access is managed via Google Cloud projects with commercial and non-commercial options.\n"],null,["# About Google Earth Engine\n\nGoogle Earth Engine is a [Google Cloud product](https://cloud.google.com/earth-engine) for geospatial\nanalysis at scale. It combines a multi-petabyte catalog of satellite imagery and\ngeospatial datasets with planetary-scale computation to accelerate environmental\nresearch and applications.\n\nKey Features\n------------\n\n### Geospatial analysis, simplified and scalable\n\nEarth Engine integrates an extensive geospatial [data\ncatalog](/earth-engine/datasets) with distributed computing, accessible through\nclient libraries. Users can access a wide range of satellite and environmental\ndata, as well as [incorporate their own datasets](/earth-engine/guides/image_upload). The platform\nsimplifies geospatial analysis by automatically handling data projection,\nscaling, and compositing based on user-specified parameters. Its [analytical\nfunctions](/earth-engine/guides/objects_methods_overview) operate efficiently across different scales without\nrequiring explicit data preparation steps or chunking. By managing complex data\nprocessing and computational scaling internally, Earth Engine enables users to\nfocus on analysis rather than technical setup.\n\n### Processing environments\n\nEarth Engine supports [two modes of analysis](/earth-engine/guides/processing_environments):\n\n- **Interactive mode**: For rapid real-time data exploration and visualization of small amounts of data.\n- **Batch mode**: For large-scale computationally intensive tasks on large amounts of data.\n\n### Development environments\n\nDevelopers can choose between two primary development environments:\n\n- **Python client library**: A flexible interface to Earth Engine for integration with the broader Python ecosystem, facilitating advanced workflows, and interactive analysis in Jupyter notebooks.\n- **JavaScript Code Editor**: A dedicated web-based development environment for rapid prototyping, exploration, and Earth Engine App creation.\n\n### Visualization and results\n\nEarth Engine supports geospatial analysis from initial prototyping to final data\nexport. Its efficient tiling and computation system, integrated with interactive\nmap widgets, provides rapid visualization and inspection capabilities in both\nthe Code Editor and Python environments. This allows for immediate data\nexploration and iteration. When ready, users can [export](/earth-engine/guides/exporting) raster\nand vector results to Google Cloud Storage, BigQuery, or Google Drive, as well\nas download data locally in formats compatible with pandas, NumPy, and Xarray.\nAdditionally, Earth Engine supports the creation of [interactive web\napplications](/earth-engine/guides/apps), enabling users to share their geospatial insights with\na wide audience.\n\n### Machine learning\n\n[Machine learning tools](/earth-engine/guides/machine-learning) for regression, classification, image\nsegmentation, and accuracy assessment are built into Earth Engine. Once trained,\nmodels can be saved and applied repeatedly. Classical ML workflows are\nstreamlined within Earth Engine's integrated system. For more advanced options\nor externally trained models, integration with [Vertex AI](https://cloud.google.com/vertex-ai/docs/start/introduction-unified-platform) is\nprovided, allowing models to be brought to Earth Engine's data or enabling the\nconstruction of deep learning models and neural network-based analyses.\n\nAccess and management\n---------------------\n\nEarth Engine is available for both [commercial](https://earthengine.google.com/commercial/) and\n[noncommercial](https://earthengine.google.com/noncommercial/) use. Noncommercial use is offered free of\ncharge, while commercial use is subject to a [subscription fee and compute\ncharges](https://cloud.google.com/earth-engine/pricing). All computation and private data are associated with Google\nCloud projects, providing users with control over access, resource management,\nand usage monitoring through the Google Cloud Console. This integration allows\nfor centralized project management, detailed billing information, and the\napplication of Google Cloud's robust security and compliance features. Users can\ntake advantage of Identity and Access Management (IAM) to [control\npermissions](/earth-engine/cloud/access-control) and can [log activities](/earth-engine/guides/audit_logging) and [monitor\nresource usage](/earth-engine/guides/monitoring_usage) with Cloud Monitoring and Cloud Logging."]]
