geospatial
Satellite Annotation
for Waste Site
Detection
United States
Overview
Overview
A US-based academic researcher needed reliable training data to detect informal landfills in satellite imagery. Label Your Data managed independent site reviews and resolved conflicting assessments to build a consistent dataset for the researcher’s computer vision model.
reviewed
reviews
team scaled
Client
A US-based academic researcher was developing a comprehensive dataset of informal landfill locations using satellite imagery. The research aimed to support automated waste-site detection through computer vision, requiring reliable labels for locations that could be difficult to identify through visual inspection alone.
Challenges
With no suitable labeled dataset available for the project, identifying informal landfills required extensive manual satellite imagery review. Subtle visual differences made some waste sites difficult to recognize, increasing the risk of false positives and missed detections.
The researcher needed a repeatable process for verifying large volumes of locations while maintaining consistent labeling decisions.
Solution
Label Your Data used Google Earth to inspect potential landfill locations and established an independent review workflow. Initial onboarding covered false positives and false negatives, while the account manager trained additional team members using updated guidelines and edge-case examples as the project scaled.
Independent dual review, with two annotators assessing each location
Disagreement resolution before finalizing annotations
Centralized satellite imagery annotation tracking in shared Google Sheets
Flexible team scaling from 2 to 10 annotators, accommodating project pauses and workload changes
Need Reliable Ground Truth Data for
Complex Satellite Imagery?
Results
The review process provided the researcher with a consistently checked dataset for informal landfill detection. Following annotation, the model achieved a substantially higher ROC-AUC score, while the team completed the validation workload within the project’s delivery timeline.
40,000 locations reviewed twice, totaling 80,000 independent assessments
ROC-AUC improved from 0.70 to 0.96 after dataset labeling
Final batch delivered in about two months, with no backlog
Why AI Teams Choose Label Your Data
Data Annotation for Complex Environments
Rely on consistent, high-quality output for complex datasets, detailed taxonomies, and edge cases.
Structured Quality from Pilot to Production
Get quality engineered into every step through onboarding, evolving guidelines, QA, and continuous feedback.
Flexible and Scalable Operations
Adjust team capacity, project size, and delivery model as you scale, with no setup fees or long-term lock-ins.
An Integrated Delivery Partner
Align on goals, workflows, and expectations with a team that integrates into your process from day one.
Projects Led by Annotation Experts
Work with former annotators who understand annotation complexity, quality standards, and high-volume delivery.
Request a pilot
Tell us more about your project and data
Thank you for contacting us!
We'll get back to you shortly
Fast, responsive, accurate, and flexible. Label Your Data has been great to work with and consistently delivers exactly what we need.
Academic Researcher
US-Based Satellite Imagery Research Project
Trusted by ML Professionals
Explore More Case Studies
Aerial Image Annotation for Land Cover Classification
Land Segmentation for Drone Reforestation
