2026-09-03 –, Dahlia1
In this session, we will discuss cloud-native drone processing pipeline using OpenDroneMap (ODM). Learn how to reliably transform standard 2D drone imagery into high-resolution Ortho mosaics, 3D point clouds, and elevation models, balancing automated quality assurance with manual cloud cost controls.
Managing massive drone datasets from the field to final geospatial products requires robust orchestration. To solve this, our team developed a custom, cost-efficient cloud processing pipeline. We will explain exactly how this architecture worked for us, showing how standard 2D drone imagery is transformed into high resolution ortho mosaic (basemap) rich, measurable 3D outputs and elevation models using open-source tools.
We will discuss the photogrammetry pipeline, focusing on practical implementation with OpenDroneMap (ODM).
Structure from Motion (SfM): Understanding the core concept how algorithms calculate camera positions and match features in overlapping JPEGs to extract dense 3D point clouds.
OpenDroneMap & Docker: A quick overview of what ODM is, how to deploy containerized Docker ODM for reproducible environments, and a guide to the most important processing flags.
Elevation & Canopy Modeling: Moving beyond flat Ortho mosaics to generate Digital Surface Models (DSM), bare-earth Digital Elevation Models (DEM), and Canopy Height Models (CHM) for advanced vegetation analysis.
Hardware & Cloud Compute: Photogrammetry requires significant computing power. We will explain how AWS services helped in our implementation.
In this session for any geospatial professionals predominantly working with the macroscopic scale of satellite imagery, this session will offer an practical introduction to the ultra-high-resolution capabilities of drone data processing.
OpenDroneMap (ODM): The core photogrammetry engine of our pipeline. It is essential for executing the Structure from Motion (SfM) algorithms that process raw 2D drone images into 3D point clouds, orthomosaics, and elevation models.
Docker: Essential for deploying containerized OpenDroneMap (Docker ODM). It ensures a reproducible and scalable environment to run heavy photogrammetry processing reliably in the cloud.
QGIS: Essential for visualizing and analyzing the generated 2D geospatial assets, such as the high-resolution orthomosaics, Digital Elevation Models (DEM), and Canopy Height Models (CHM).
CloudCompare: Essential for rendering, inspecting, and visualizing the dense 3D point clouds generated by ODM, as it provides specialized, high-performance tools for 3D spatial data.
I make my conference contribution available under the CC BY 4.0 license. The conference contribution comprises the abstract, the text contribution for the conference proceedings, the presentation materials as well as the video recording and live transmission of the presentation:
I am a geospatial data science professional specializing in remote sensing, machine learning, and scalable workflows. I currently manage government geospatial R&D initiatives focused on environmental monitoring and land management.
As co-founder of the "Let's Talk Spatial" meetup, we work to make spatial concepts accessible to wider audiences. I am driven by the intersection of geospatial intelligence, AI, and strategy to build scalable solutions that create real-world environmental impact.