Intro to OpenDroneMap

Center for Geospatial Analytics at North Carolina State University

Corey T. White

Overview

  • Why OpenDroneMap?
  • ODM architecture and pipeline
  • WebODM interface
  • Comparison with Agisoft Metashape
  • ODM + GRASS integration

Why OpenDroneMap?

  • Open-source photogrammetry ecosystem
  • Replaces commercial tools (Agisoft, Pix4D) in many workflows
  • Ideal for teaching, research, reproducibility
  • Actively maintained
  • Community-driven (UAV4GEO, Public Lab, HOTOSM)

The Open-Source Photogrammetry Landscape

Commercial

  • Agisoft Metashape
  • Pix4D
  • Correlator3D

Open-Source

  • OpenDroneMap (ODM)
  • WebODM
  • NodeODM
  • MicMac

What is OpenDroneMap?

ODM Ecosystem

  • ODM is a command-line photogrammetry processing engine (OpenDroneMap Authors 2020).
  • NodeODM is a REST API to access ODM.
  • WebODM is a Web/GUI interface for ODM.
  • CloudODM is a command line client to work with ODM with NodeODM.
  • PyODM is a Python SDK for scripting with ODM.
  • ClusterODM is a NodeODM API compatible reverse proxy, load balancer and task tracker for easy horizontal scaling.
  • NodeMICMAC is a Node.js App and REST API to access MicMac.
  • FIELDimage - is a Tool to Analyze Images From Agricultural Field Trials and Lab in R (Matias, Caraza-Harter, and Endelman (2020), Pawar and Matias (2023)).
  • Find-GCP - Find ArUco markers in digital photos (Siki and Takács (2021)).

ODM Architecture

  • ODM Command-line engine (OpenDroneMap Authors (2020))
  • Built from modular components:
    • OpenSfM – SfM + Bundle Adjustment
    • openMVS – Dense reconstruction
    • MicMac – (Optional) photogrammetry processing engine
    • PDAL – Point cloud processing (PDAL Contributors (2025))
    • Entwine – Point cloud data organization library
    • GDAL – Raster operations
    • PoissonRecon – 3D mesh
    • MSV Texturing – Mesh Texturing
    • GRASS – geospatial compuational engine (GRASS Development Team et al. (2025))

Platform and Hardware Requirements

Platforms

Through Docker ODM, NodeODM, and WebODM run on:

  • Windows
  • MacOS
  • Linux

Note

Docker is an open platform for developing, shipping, and running applications. For more information visit What is Docker?.

Recommendations from (Toffanin (2023))

Hardware Requirements

Minimum requirements

  • 64 bit CPU (manufactured post-2010)
  • 20 GB of disk space
  • 4 GB of RAM

Can process no more than 100-200 images.

  • Latest generation CPU
  • 100 GB disk space
  • 16 GB RAM

Note

NVIDIA GPU can be used to speed up processing.

Recommendations from (Toffanin 2023)

ODM Capabilities

  • Feature detection (SIFT-like)
  • Bundle adjustment
  • Dense cloud generation
  • DSM & DTM creation
  • Orthomosaic generation
  • Meshing & texturing
  • Ground/non-ground classification
  • QC & reports
  • GCP support

ODM Processing Workflow

flowchart LR
    subgraph Client
        direction LR
        A[Import Images]-->B[Quality Check]
        B -.-> GCP["(Optional) GCP"]
    end
    subgraph SfM
     direction TB
        GCP --> C[Feature Detection & Matching]
        B --> C[Feature Detection & Matching]
        C --> D[Bundle Adjustment]
        D --> E[Dense Point Cloud]
    end
    subgraph Processing
     direction BT
        E --> F[DSM/DTM Generation]
        F --> G[Orthomosaic]
        G --> H[Export Outputs]
    end

WebODM

WebODM Overview

  • Browser-based interface to ODM
  • Easy installation via Docker
  • Project-based workflow
  • Supports GCP import
  • Built-in 2D and 3D viewers
  • Plugin ecosystem (Contours, Volume, Reports)

Basic WebODM Workflow

  1. Create project
  2. Upload images
  3. (Optional) Upload GCP file
  4. Run processing
  5. Inspect outputs
  6. Export and analyze in GRASS

WebODM Dashboard

WebODM - New Project

Upload Images

(Optional) Create GCP file

Run Process

Inspect outputs 2D

Map View

Inspect outputs 3D

Potree Viewer

Export Assets

ODM vs Agisoft Metashape

Strengths of ODM

  • 100% open-source
  • Transparent algorithms
  • Strong reproducibility
  • Easy scripting/automation
  • Works well with GRASS/QGIS
  • Excellent orthomosaics for natural scenes

Limitations of ODM

  • Slower than Metashape
  • Mesh reconstruction less refined
  • GUI less polished (WebODM improves this)
  • Multispectral workflows less robust
  • Struggles with highly oblique imagery

Comparison Table

Capability ODM WebODM Agisoft
Cost Free Free $$$
Orthomosaic Excellent Excellent Excellent
Dense Cloud Quality Good Good Excellent
GUI Basic CLI Browser GUI Professional
Speed Slower Slower Very fast
Scripting Excellent Good Good
GCP Handling Good Good Excellent

References

GRASS Development Team, Martin Landa, Markus Neteler, Markus Metz, Anna Petrášová, Vaclav Petráš, Glynn Clements, et al. 2025. “GRASS GIS.” https://doi.org/10.5281/zenodo.4621728.
Matias, Filipe Inácio, Maria V. Caraza-Harter, and Jeffrey B. Endelman. 2020. “FIELDimageR: An r Package to Analyze Orthomosaic Images from Agricultural Field Trials.” The Plant Phenome Journal 3 (1): e20005. https://doi.org/10.1002/ppj2.20005.
OpenDroneMap Authors. 2020. “ODM: A command line toolkit to generate maps, point clouds, 3D models and DEMs from drone, balloon or kite images.” GitHub repository. https://github.com/OpenDroneMap/ODM.
Pawar, Popat S., and Filipe Inacio Matias. 2023. “FIELDimageR.extra: Advancing User Experience and Computational Efficiency for Analysis of Orthomosaic from Agricultural Field Trials.” The Plant Phenome Journal 6 (1): e20083. https://doi.org/10.1002/ppj2.20083.
PDAL Contributors. 2025. “PDAL Point Data Abstraction Library.” https://doi.org/10.5281/zenodo.10884408.
Siki, Zoltan, and Bence Takács. 2021. “Automatic Recognition of ArUco Codes in Land Surveying Tasks.” Baltic Journal of Modern Computing 9 (1). https://doi.org/10.22364/bjmc.2021.9.1.06.
Toffanin, Piero. 2023. OpenDroneMap: The Missing Guide. Second. UAV4GEO.