Helping Engineers Locate Walls Using GIS & AI
Skill Level: Beginner
Intended Audience: GIS User, GIS Developer, Student
Abstract
As GIS professionals, we’re hearing more about artificial intelligence, but many organizations are still determining how to apply it to everyday projects. Horner & Shifrin developed and trained a computer vision model to identify Mechanically Stabilized Earth (MSE) walls from aerial imagery. Starting with manually identified examples, GIS staff created training datasets, reviewed model outputs, and iteratively improved accuracy through QA/QC and validation. The process highlighted an important reality of AI in GIS: successful models depend on geospatial data and GIS professionals. Beyond the technical workflow, this project demonstrates how GIS professionals can lead AI initiatives by combining spatial analysis, data management, and human validation. Attendees can learn practical lessons on training AI models, creating feedback loops to improve performance, and integrating AI into existing GIS workflows. The presentation offers a realistic example of how AI can enhance efficiency while reinforcing the critical role of GIS professionals in developing reliable, production-ready solutions.
Presenter
Elise Bush
GIS technician, Horner & Shifrin
Graduated from Western Illinois University with MS in GIScience and Geoenvironment in 2024. Had an internship with the Office of Geospatial Information at the State of Missouri in 2024. Began working as a GIS technician at Horner & Shifrin in March of 2025.