Satellite & Panorama Annotation
Budget: £20 – £250 GBP
I need help expanding an internal computer-vision dataset by carefully annotating both satellite and street-level panoramic imagery. The focus for this phase is the precise outlining and tagging of Buildings only; roads, vegetation, and other elements are not part of the brief right now, so every polygon or bounding box you draw must relate to man-made structures.
You’ll be working in a web-based labeling environment that exports standard formats (COCO, GeoJSON), so basic software-engineering literacy and an eye for organized file structures will make the hand-off smoother. Familiarity with geospatial concepts—map projections, scale variation between zoom levels, parallax in panoramas—will help you keep labels consistent across different tile sizes and camera angles.
When you apply, highlight your relevant experience: previous large-scale labeling projects, GIS or remote-sensing coursework, or any pipeline automation you have built around annotation tasks. No lengthy proposal is needed; a concise summary of what you have done and links or screenshots of similar work are enough.
Deliverables
• Fully annotated image batches returned in the requested format
• A brief log explaining any edge cases or uncertainties encountered
I will review a small pilot submission for accuracy and consistency before green-lighting the full run, and I’m happy to iterate on the labeling guidelines with you as we go.
You’ll be working in a web-based labeling environment that exports standard formats (COCO, GeoJSON), so basic software-engineering literacy and an eye for organized file structures will make the hand-off smoother. Familiarity with geospatial concepts—map projections, scale variation between zoom levels, parallax in panoramas—will help you keep labels consistent across different tile sizes and camera angles.
When you apply, highlight your relevant experience: previous large-scale labeling projects, GIS or remote-sensing coursework, or any pipeline automation you have built around annotation tasks. No lengthy proposal is needed; a concise summary of what you have done and links or screenshots of similar work are enough.
Deliverables
• Fully annotated image batches returned in the requested format
• A brief log explaining any edge cases or uncertainties encountered
I will review a small pilot submission for accuracy and consistency before green-lighting the full run, and I’m happy to iterate on the labeling guidelines with you as we go.