The Taylor Geospatial Engine (TGE) Innovation Bridge program represents a novel approach to commercializing academic research in geospatial technology. By strategically connecting academic researchers with geospatial product developers and industry stakeholders, TGE addresses a critical bottleneck: insufficient time, talent, and resources to transform groundbreaking geospatial research into practical commercial applications.
This case study examines TGE’s Innovation Bridge program in action and how it produced Fields of the World (FTW), now the largest global benchmark dataset for agricultural field boundary detection with an ecosystem of models, tools and output datasets. The outcomes from FTW demonstrate how the Innovation Bridge program can transform academic innovation into industry-ready solutions, while maintaining open source principles and enabling broad commercial adoption.
The Challenge and Solution
Academic researchers are regularly pushing the state of the art forward in geospatial science, particularly in applying artificial intelligence and machine learning to Earth observation data. These innovations have the potential to unlock new opportunities for industry to meet customer needs at scale.
However, the work is often done across labs without the funding or remit to ensure that a breakthrough can unlock further innovation quickly. The obstacles to broader adoption are often format incompatibility, data stored in ways that make it difficult to discover or access effectively, and lack of industry-specific expertise in development. These barriers result in a high cost of validation for commercial use and inhibit follow-on innovation.
TGE identified a drag in geospatial innovation during a presentation by Dr. Hannah Kerner when she asked the question “where is my imagenet trained resnet for remote sensing?”
While mainstream computer vision has abundant labelled training data for standard photographs, as a result of Fei Fei Li’s labor of love building imagenet, very little training data exists for satellite imagery. Li’s work on ImageNet was a valiant effort to create a shared data resource that the entire research community could build upon and it ultimately reshaped the industry.
“In order to speed up the development and commercialization of geospatial innovation emerging from academic research, we need to unlock bottlenecks in ways that create a springboard for future innovation”
– Jennifer Marcus, TGE Executive Director
Progress on GeoAI is lagging: to break past the need for manual human review of satellite imagery, there is a need for an open ecosystem around which to innovate and share progress. TGE’s purpose is to address these issues by running a series of Innovation Bridge programs designed to unlock an entirely new generation of geospatial breakthroughs.
Bridging Academia and Industry
TGE addresses the geospatial innovation gap by bringing researchers and industry together, embedding industry-leading software developers alongside academics during the research process itself. This collaboration ensures that as papers are published, the accompanying tools and outputs are already aligned with industry needs, accessible to broad audiences, and ready for immediate use. By integrating commercial considerations at the research stage rather than afterward, TGE dramatically reduces the work required for practitioners to benefit from academic breakthroughs.
The Innovation Bridge is unique in its approach. It preserves academic integrity by allowing researchers to maintain their focus on scientific advancement while TGE handles commercialization. A variety of stakeholder teams ensure collaborative development with diverse perspectives, while an open-by-design philosophy publishes foundational results openly to enable broad adoption, leaving room for proprietary innovation to be built.
Innovation Bridge in Practice:
Fields of The World (FTW) TGE identified that field boundary extraction from satellite imagery was a foundational challenge to solve. Field boundaries are key datasets that enable sophisticated insights for agricultural monitoring, precision farming, carbon measurement, climate regulation compliance, land use evaluation, food security assessments, and deforestation monitoring.
While governments in some countries maintain high-quality field boundary data, most of the world lacks comprehensive coverage. Machine learning could theoretically extract this information from satellite imagery at global scale, but the community lacked diverse training datasets representing global agricultural landscapes, standardized data formats for interoperability, validation against industry requirements, and clear pathways to commercial deployment.
From Concept to Reality: Convening in St.Louis
TGE organized a groundbreaking two-day workshop in St. Louis that embodied the Innovation Bridge philosophy by bringing together stakeholders from academia, industry, and government. The gathering included representation from Microsoft’s AI for Good Lab, Oak Ridge National Laboratory, Washington University, World Resources Institute, and agricultural industry partners including Bayer and Planet.
The group tackled critical questions about technical specifications, data standards, and licensing considerations. Industry partners shared their actual workflows and pain points, transforming abstract possibilities into concrete requirements, while researchers contemplated how to push the state of the art in ML forward around this problem. The workshop resulted in not just a project plan, but a shared vision of success from multiple perspectives—and a commitment from all stakeholders to advance agricultural AI applications that no single organization could have achieved alone.
“This project brings together the dream team in a way I’ve never seen before—you have academic professors who are wizards at modeling, industry experts who know how to deploy models and handle geospatial data efficiently at scale, and anyone who’s anybody in the field is just one ping away from joining the conversation. Other projects I’ve been involved in don’t bring the right people together like this; here we have a full squad with all the expertise we need.”
– Isaac Corley, Sr. Machine Learning Engineer at Wherobots and FTW contributor
Demonstrating Value Through Results
The Innovation Bridge has generated significant outcomes through the Fields of The World (FTW) initiative. The project has advanced academic research through peer-reviewed publications and benchmark datasets, while achieving substantial technical performance improvements in model accuracy and processing speed.
Beyond research advances, FTW has enabled commercial adoption by major organizations addressing real-world sustainability challenges and transformed industry practices by reducing development costs and barriers to entry. These interconnected results illustrate how bridging academia and industry creates value that extends far beyond any single stakeholder.
A New Model for Geospatial Innovation
The Innovation Bridge addresses a systemic challenge where promising research fails to reach practical application. TGE builds literal bridges in the form of connections, infrastructure, standards, and communities. This approach accelerates the pace at which geospatial innovation can create real-world impact.
As the geospatial industry faces growing demand for AI-powered Earth observation capabilities, the Innovation Bridge framework offers a blueprint for how academic research, industry needs, and open collaboration can combine to unlock value that benefits researchers, companies, and society. The Fields of the World initiative proves the concept. The question now is how many more bridges can be built.
Find out more at www.fieldsofthe.world
TGE kindly acknowledges all the participants who have been involved in contributing to Fields of The World: Arizona State University, Bayer’s Ecosystem Services team, Clark University, Danforth Plant Science Center, Microsoft’s AI for Good Lab, Miraterra, Oak Ridge National Laboratory (ORNL), Oregon State University, Planet Labs, Radiant Earth, Taylor Geospatial Institute, UN Forestry and Agriculture Organization, Washington University, Wherobots, World Resources Institute (WRI). We also wish to recognize the extremely talented cadre of Technical Fellows led by our Senior Strategist Chris Holmes: Jeff Albrecht, Ivor Bosloper, Isaac Corley, Andreas Hocevar, Matthias Mohr, Martha Morrissey, Zach Richardet, and Andrew Smith.

