Mastering AI Workflows with FME by Mark Döring
0 likes•112 views
Harness the full potential of AI with FME: From creating high-quality training data to optimizing models and utilizing results, FME supports every step of your AI workflow. Seamlessly integrate a wide range of models, including those for data enhancement, forecasting, image and object recognition, and large language models. Customize AI models to meet your exact needs with FME’s powerful tools for training, optimization, and seamless integration
Mastering AI Workflows with FME by Mark Döring
- 1. Mastering AI Workflows with FME con terra GmbH The Peak of Data and AI 2025
- 2. 2025 The Peak of Data and AI Mark Döring Head of Data Integration Project Services con terra GmbH
- 3. 1. Introduction 2. Geo AI at con terra 3. General Approach 4. Project Example KINoPro 5. Conclusion Agenda
- 4. Geo:AI is a Machine Learning technology, which enables the caption and analysis of complex patterns and structures in spatial data.
- 7. Geo:AI at con terra
- 8. 2025 The Peak of Data and AI Setup of a „typical“ Geo:AI Project Data Pre-processing AI-Methods Integration and Operation Visualization and Application
- 9. 2025 The Peak of Data and AI Integration in existing processes and infrastructures
- 10. 2025 The Peak of Data and AI Setting it all together • GeoAI Connector for FME • Connecting to an external Python interpreter • No version conflicts of Python modules • Maintenance of AI runtime is separate from FME • Use the AI Model within FME Workflows
- 15. 2025 The Peak of Data and AI Project - KINoPro Künstliche Intelligenz zur NonnenfalterPrognose ● Research Project of TU Dresden and con terra GmbH ● Data from state forestries Brandenburg and Saxony ● Climate Change Influence ● Trees struggling with dry and hot weather ● Forest pests adapt faster than plants ● Irregular population growth/appearence ● Forestry personnel needs to be managed more efficiently ● New prediction models are necessary
- 17. Approach - Overview | Monitoring of Nun Moth / Black Arches using traps GeoAI-Mod ell Influencing Factors Data Research Web Application Data Preparation and Preprocessing Validation and survey for additional parameters
- 18. Predicting insect populations Challenges Prediction made more difficult by climate change Diverse parameters ○ land cover classification, elevation, slope, trapposition and orientation, air and soil moisture… ○ 6,000 data points, data from 70 weeks per point ○ Heterogeneous data formats (Excel, ASCII grids, GeoTIFFs, netCDF ...) ○ Spatial correlation of traps is critical
- 19. Data Preprocessing with FME ASCII Grids netCDF Gridded Binary Esri Geodatabase AI-Algorithms Data Preprocessing Generating Geo Data
- 20. Data Preprocessing with FME ASCII Grids netCDF Gridded Binary Esri Geodatabase AI-Algorithms Data Preprocessing Generating Geo Data
- 21. Easy-to-Use Integration of ML-Algorithms Library Algorithms Custom Transformers
- 22. Learnings ● Artificial neural network learns correlations between specialized data, weather influences and previous year's data ● FME can predict insects populations approx. 5 months in advance ● Identification of high-risk regions and arches populations ● Reduction of traps and less traffic in the forest ● Less chemical distribution and use ● Optimized operational planning
- 23. 2025 The Peak of Data and AI Recap: Harnessing Data Integration Use this slide to: ● Reaffirm your objective ● What the audience learned ● Outline next steps if any
- 24. 2025 The Peak of Data and AI “FME was an AI Playground and becomes a Geo:AI Solution”
- 26. 2025 The Peak of Data and AI “You can use this slide to include a quote as well.” — Attribution Name, Relevant Title
- 27. 2025 The Peak of Data and AI Recap: Harnessing Data Integration Use this slide to: ● Reaffirm your objective ● What the audience learned ● Outline next steps if any