Case Study:
Taking an urban tree-monitoring AI from research to operational rollout
Sector:
Urban forestry / Municipal
BaKIM / Smart City Bamberg
Scale:
City-wide (Bamberg, Germany, expanding to other municipalities)
Service pillars:
Precision Inventory · Sensor Strategy
Smart City Bamberg develops "digital technologies designed to make Bamberg an even more livable place. [Their] focus is on people, their needs, and their desires for a connected, innovative, and vibrant Bamberg.”
Their BaKIM project seeks to use AI to help care for the city’s trees and green spaces.
The Brief:
Cities across Europe are facing a dual pressure: urban tree canopies are declining due to heat stress, drought, pests, and disease, while the manual surveys needed to monitor them are expensive and slow. A trained arborist can typically assess around 150 trees per day. Bamberg alone has thousands of municipal trees, and the City needed a way to survey them at scale, detect health issues early, and prioritise maintenance before problems escalate.
The BaKIM project (“Baum, KI, Mensch”: Tree, AI, Human), led by Smart City Bamberg with the University of Bamberg’s Chair of Cognitive Systems, set out to build an AI-powered web application that would let arborists and foresters upload drone imagery and receive automated tree inventory and health assessments. The underlying computer-vision models were developed at the University of Bamberg, with Jonas Troles as technical lead and Prof. Dr. Ute Schmid leading the accompanying research.
What forestmap.ai delivered:
We were brought in as the AI and computer-vision experts to help transition from research to operational use. Our role focused on the move from prototype to production:
Advising on the computer-vision approach, drawing on our individual-tree detection work (including detectree2) and tropical-forest experience
Optimisation of the detection and classification pipeline for reliable, repeatable performance on real municipal drone imagery
Enhanced capabilities and refinements that extended what the system could do in the field - tree crown delineation, species classification, and vitality assessment for issues such as bark-beetle and mistletoe infestation and drought stress
Helping integrate the models into a deployable web application that municipal staff can actually use
The resulting application lets municipal staff upload drone footage, have it processed server-side by the AI, and download GIS-compatible results for their tree-management workflows.
Ready to discuss your forest?
Reach out to find out more about how forestmap.ai can help your organization.
info@forestmap.ai
+44 (0)7986 882 740
By submitting this form, you acknowledge our Privacy Policy.
