Metashape Azure

Installing

# cmd

conda create --name sfm python=3.8 -y
conda activate sfm

pip install uv 

uv pip install "git+https://github.com/Jordan-Pierce/Metashape-Azure.git"

Fallback: if you get an error related to --prerelease=allow, run the command below

uv pip install "git+https://github.com/Jordan-Pierce/Metashape-Azure.git" --prerelease=allow

Finally, to run the application, use the following command:

# cmd

metashape-azure-mls

Updating

# cmd

conda activate sfm

uv pip install -U "git+https://github.com/Jordan-Pierce/Metashape-Azure.git"

Running the Application

1. Open Azure Machine Learning Studio in Microsoft Edge and login
2. Navigate to your workspace
3. Copy the following credentials from the top-right of the page:
   • Subscription ID
   • Resource Group
   • Workspace Name
4. Enter credentials in the application:
   • Use "Save Credentials" to store for future use
   • Click "Authenticate"
5. Select your compute cluster from the dropdown
6. Configure project paths:
   • Input Path: Azure URI to image folder
   • Output Path: Azure URI for project destination
   • Project Name: Unique name for your project

     Note: {output path}/{project name} must not already exist

7. Select desired SfM parameters
8. Click "Run on Azure" to start processing

Notes

• Azure Machine Learning Studio and Authentication must be done in Microsoft Edge
• You must be connected to the network (i.e., VPN) to access the Azure services.
• Currently using python 3.8, and Metashape 2.2.3
• See here for older versions of Metashape Python API wheels

Linux environments

To ensure you can run Metashape on a Linux distro (e.g., Ubuntu), please ensure you do the following:

echo 'export AGISOFT_FLS=10.71.68.13:5842' >> ~/.bashrc && source ~/.bashrc

sudo apt install libglu1-mesa libgl1-mesa-glx libcurl4

Then you should be able to run python in the terminal, import Metashape and see that the license was pulled.