使用地理人工智能数据科学虚拟机Using the Geo Artificial Intelligence Data Science Virtual Machine

使用地理 AI 数据科学 VM 提取数据进行分析、执行数据争论,以及为使用地理空间信息的 AI 应用程序构建模型。Use the Geo AI Data Science VM to fetch data for analysis, perform data wrangling, and build models for AI applications that consume geospatial information. 在预配地理人工智能 Data Science VM 并通过 ArcGIS 帐户登录到 ArcGIS Pro 后,可以开始在线与 ArcGIS 桌面和 ArcGiS 进行交互。After you've provisioned your Geo AI Data Science VM and signed in to ArcGIS Pro through your ArcGIS account, you can start interacting with ArcGIS desktop and ArcGIs online. 还可以通过地理 Data Science VM 上预配置的 Python 接口和 R 语言桥来访问 ArcGIS。You can also access ArcGIS from Python interfaces and an R language bridge that's preconfigured on the Geo-Data Science VM. 若要构建丰富的 AI 应用程序,请将地理 Data Science VM 与在其上可用的机器学习和深度学习框架以及其他数据科学软件组合使用。To build rich AI applications, combine the Geo-Data Science VM with the machine-learning and deep-learning frameworks and other data science software that are available on it.

配置详细信息Configuration details

Python 库 arcpy,用于与在数据科学 VM 的全局根 conda 环境(位于 c:\anaconda)中安装的 ArcGIS 进行交互。The Python library, arcpy, which is used to interface with ArcGIS, is installed in the global root conda environment of the Data Science VM that's found at c:\anaconda.

  • 如果在命令提示符运行 Python,请运行 activate 来激活到 conda 根 Python 环境。If you're running Python at a command prompt, run activate to activate into the conda root Python environment.
  • 如果使用 IDE 或 Jupyter Notebook,可以选择环境或内核来确保处于正确的 conda 环境中。If you're using an IDE or Jupyter notebook, you can select the environment or kernel to make sure you're in the correct conda environment.

ArcGIS 的 R 桥作为名为 arcgisbinding 的 R 库安装在位于 C:\Program Files\Microsoft\ML Server\R_SERVER 的主 Microsoft Machine Learning Server 独立实例中。The R bridge to ArcGIS is installed as an R library named arcgisbinding in the main Microsoft Machine Learning Server standalone instance that's located at C:\Program Files\Microsoft\ML Server\R_SERVER. Visual Studio、RStudio 和 Jupyter 已预配置为使用此 R 环境并将具有对 arcgisbinding R 库的访问权限。Visual Studio, RStudio, and Jupyter are already preconfigured to use this R environment and will have access to the arcgisbinding R library.

地理 AI 数据科学 VM 示例Geo AI Data Science VM samples

除了基础 Data Science VM 提供的基于机器学习和深度学习框架的示例,地理人工智能 Data Science VM 还提供了一组地理空间示例。In addition to the machine-learning and deep-learning framework-based samples from the base Data Science VM, a set of geospatial samples is also provided as part of the Geo AI Data Science VM. 这些示例可以帮助你使用地理空间数据和 ArcGIS 软件快速开始 AI 应用程序开发:These samples can help you jump-start your development of AI applications by using geospatial data and the ArcGIS software:

  1. 开始使用 Python 地理空间分析:一个介绍性示例,介绍了如何使用 arcpy 库提供的到 ArcGIS 的 Python 接口来处理地理空间数据。Getting started with geospatial analytics with Python: An introductory sample showing how to work with geospatial data through the Python interface to ArcGIS is provided by the arcpy library. 它还介绍了如何将传统的机器学习与地理空间数据组合使用,并在 ArcGIS 中的地图上直观呈现结果。It also shows how to combine traditional machine learning with geospatial data and then visualize the result on a map in ArcGIS.

  2. 使用 R 进行地理空间分析入门:一个介绍性示例,介绍了如何使用 arcgisbinding 库提供的到 ArcGIS 的 R 接口来处理地理空间数据。Getting started with geospatial analytics with R: An introductory sample that shows how to work with geospatial data by using the R interface to ArcGIS that's provided by the arcgisbinding library.

  3. 像素级土地利用分类:该教程展示了如何创建深度神经网络模型,使该模型接受航拍图像作为输入并返回土地覆盖标签。Pixel-level land use classification: A tutorial that illustrates how to create a deep neural network model that accepts an aerial image as input and returns a land-cover label. 土地覆盖标签的示例包括“森林”和“水” 。Examples of land-cover labels are forested and water. 模型会为图像中的每个像素返回这样的标签。The model returns such a label for every pixel in the image.

后续步骤Next steps

使用数据科学 VM 的其他示例位于以下位置:Additional samples that use the Data Science VM are available here: