Jump to content

Fachhochschule Dortmund

Language

SmartFarmScape

Entwicklung eines mobilen, heterogenen Sensornetzerkds, bestehend aus einem autonom gesteuerten Drohnenschwarm von mindestens 20 Drohnen, zur Überwachung landwirschaftlicher Flächen und zur Analyse urbaner Hitzeinseln

Fast facts

  • Organizational unit

  • Category

    • Federal project
    • Foundations etc.
  • Funding source

    Federal Ministry of Business Studies and Climate Protection - BMWK,unknown organization

  • Funding program

    BMWK Central Innovation Program for SMEs (ZIM), third-party funds

  • Duration

About the project

Project plans

Development of an AI-supported, decentralized multi-agent drone system for large-scale environmental monitoring using panoramic imaging and time series analysis

Project motivation

In light of the increasing need for precise environmental monitoring, for example in agriculture, forestry or nature conservation, scalable, autonomous sensor networks are becoming increasingly important. Small and medium-sized enterprises in particular need flexible, cost-effective solutions for collecting and analyzing large-scale environmental data. Individual drone deployments often reach their limits in terms of area coverage, analysis automation and interoperability. There is therefore a need for intelligent systems that work efficiently, can be expanded flexibly and can also keep an eye on long-term changes.

Project description

The SmartFarmScape project is developing an innovative drone system that differs significantly from conventional agricultural drones. Instead of relying on individual, centrally controlled drones with high computing power, the project is pursuing a decentralized, scalable multi-agent approach: several low-cost drones act as a coordinated team within a self-organizing mesh network. The organization as a drone formation allows for a comprehensive collection of environmental data - more efficient, more flexible and more robust than previous solutions.

The central innovation is drone-to-drone communication, which allows control and task logic to be distributed dynamically within the network. This not only allows the system to continue working if individual drones fail, but also scales flexibly with the number of devices used. Complex calculations, such as for panorama creation or object recognition using AI, are outsourced to an accompanying edge device on the ground.

The entire process is divided into several stages: The individual images captured by the drones are compiled into high-resolution panoramas using feature-based stitching. A neural network then analyses these panoramas, recognizes trees, for example, records their positions and determines their condition. If such flights are repeated regularly, AI time series analysis can be used to automatically identify changes over time - such as growth, leaf loss or disease.

In contrast to conventional agricultural drones, which usually operate individually and send their data to the cloud for post-processing, SmartFarmScape enables immediate, coordinated and adaptive data collection in real time. The project thus opens up completely new possibilities for cost-efficient, AI-supported environmental monitoring - especially in remote regions or dynamic operational areas without a fixed infrastructure.

Promotion

Funding code

KK5659201FG4

Project partners

Contact & Team

This site uses cookies to ensure the functionality of the website and to collect statistical data. You can object to the statistical collection via the data protection settings (opt-out).

Settings(Opens in a new tab)