Running Agisoft Metashape in the Cloud: EC2, GPU Acceleration and NAS
News

Running Agisoft Metashape in the Cloud: EC2, GPU Acceleration and NAS

Posted on by admin

Agisoft Metashape is a powerful photogrammetry tool—but when working with thousands of high-resolution images, local hardware can quickly become a bottleneck. Running Metashape in the cloud offers scalable performance, remote access, and faster processing thanks to GPU instances and high-speed storage. In this guide, you’ll learn how to deploy Metashape on Amazon EC2, use GPU acceleration, and connect network-attached storage (NAS) for large-scale 3D reconstruction workflows.

Why Run Agisoft Metashape in the Cloud?

  • On-demand scalability: Spin up powerful GPU servers only when needed
  • Remote access: Collaborate with team members without local installs
  • Faster performance: Process data 5–10x faster on enterprise GPUs
  • Save on local hardware: No need to invest in expensive workstations

Recommended Cloud Setup: AWS EC2

Amazon Web Services (AWS) offers a reliable infrastructure to host Metashape. The EC2 service lets you choose virtual machines (instances) with CPU, RAM, and GPU resources.

1. Choose the Right EC2 Instance

  • g4dn.xlarge – g4dn.12xlarge: Budget GPU instances with NVIDIA T4 (good for medium projects)
  • g5.xlarge – g5.12xlarge: Newer NVIDIA A10G GPU, excellent for large datasets
  • p3.2xlarge / p4d.24xlarge: High-end NVIDIA V100/A100—best for ultra-heavy workloads

Start with a g5.4xlarge instance for a good balance of performance and cost (16 vCPU, 64 GB RAM, 1 A10G GPU).

2. Set Up the Instance

  • Go to EC2 Dashboard and launch a new instance
  • Choose an Amazon Machine Image (AMI): Windows Server 2022 or Ubuntu 22.04
  • Add a storage volume (at least 200–500 GB SSD)
  • Assign a public IP and enable RDP (for Windows) or SSH (for Linux)

3. Install Metashape

After accessing your cloud instance:

  • Windows: Download Metashape from the official website and run the installer
  • Linux: Use the .tar.gz package and run ./metashape.sh
  • Activate your license (Standard or Pro) or use trial

Ensure the proper NVIDIA drivers are installed and CUDA is enabled for GPU acceleration.

4. Install GPU Drivers (Linux)

On Ubuntu, run the following commands for NVIDIA driver and CUDA:

sudo apt update  
sudo apt install nvidia-driver-535  
sudo reboot

Verify with nvidia-smi. Then launch Metashape and go to Tools > Preferences > GPU to enable acceleration.

Mounting External Storage (NAS or EFS)

Photogrammetry projects often require terabytes of data storage. Use NAS or cloud-based storage to manage large datasets.

  • Amazon EFS (Elastic File System): Network-mounted and scalable, ideal for multi-user workflows
  • Amazon S3: Object storage for uploading/downloading large datasets
  • Remote NAS: Mount from on-premises Synology/TrueNAS using SMB or NFS

Linux example for NFS mount:

sudo apt install nfs-common  
sudo mount -t nfs 192.168.1.100:/mnt/data /mnt/nas

This allows you to store image datasets or export .PSX files directly to the NAS from Metashape.

Automation with Metashape + Python

Once your cloud instance is ready, you can automate processing with Metashape’s Python scripting API. This is ideal for:

  • Batch processing of drone missions
  • Server-side orthomosaic generation
  • Custom pipelines for classification or export

Scripts can be run directly in the Metashape Python console or from the terminal using metashape --python script.py.

Security and Licensing Notes

  • For commercial use, ensure each cloud VM has a valid license
  • You can use floating licenses if running multiple VMs (Pro edition only)
  • Always secure your cloud instance (firewall, strong passwords, no public RDP)

Cost Estimation

Cloud costs vary depending on instance type and storage used. Example (as of 2025):

  • g5.4xlarge (A10G GPU): ~$1.20/hour on-demand
  • 200 GB EBS SSD: ~$20/month
  • EFS Storage: ~$0.30/GB/month

Consider using Spot Instances to reduce costs by 60–70%—ideal for non-urgent batch processing.

Conclusion

Running Agisoft Metashape in the cloud unlocks high-performance 3D processing at scale. By leveraging EC2 GPU instances and integrating NAS or S3 storage, you can manage massive photogrammetry workloads without investing in expensive hardware. Whether you’re building a remote processing pipeline or looking to speed up a drone mapping project, the cloud offers flexibility, power, and cost-efficiency—especially when paired with Metashape’s scripting tools.

Ready to take your photogrammetry workflow to the cloud? Start with a GPU EC2 instance and experience the difference.