Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

NordFKB: a fine-grained benchmark dataset for geospatial AI in Norway

Sander Riisøen Jyhne, Aditya Gupta, Ben Worsley, Marianne Andersen, Ivar Oveland, Alexander Salveson Nossum

TL;DR

NordFKB presents a fine-grained benchmark dataset for geospatial AI in Norway, derived from the authoritative FKB database and spanning 36 semantic classes across seven geographically diverse areas. The dataset provides high-resolution orthophotos with per-class GeoTIFF masks and COCO-format bounding boxes, accompanied by random-area train/validation splits and rigorous manual quality control, plus a benchmarking repository with standardized evaluation protocols for semantic segmentation and object detection. This work enables reproducible, comparable research and practical mapping/planning applications, while acknowledging limitations in geographic coverage and class distribution. It also outlines future expansions to include temporal data and additional modalities (e.g., LiDAR) to broaden applicability and robustness.

Abstract

We present NordFKB, a fine-grained benchmark dataset for geospatial AI in Norway, derived from the authoritative, highly accurate, national Felles KartdataBase (FKB). The dataset contains high-resolution orthophotos paired with detailed annotations for 36 semantic classes, including both per-class binary segmentation masks in GeoTIFF format and COCO-style bounding box annotations. Data is collected from seven geographically diverse areas, ensuring variation in climate, topography, and urbanization. Only tiles containing at least one annotated object are included, and training/validation splits are created through random sampling across areas to ensure representative class and context distributions. Human expert review and quality control ensures high annotation accuracy. Alongside the dataset, we release a benchmarking repository with standardized evaluation protocols and tools for semantic segmentation and object detection, enabling reproducible and comparable research. NordFKB provides a robust foundation for advancing AI methods in mapping, land administration, and spatial planning, and paves the way for future expansions in coverage, temporal scope, and data modalities.

NordFKB: a fine-grained benchmark dataset for geospatial AI in Norway

TL;DR

NordFKB presents a fine-grained benchmark dataset for geospatial AI in Norway, derived from the authoritative FKB database and spanning 36 semantic classes across seven geographically diverse areas. The dataset provides high-resolution orthophotos with per-class GeoTIFF masks and COCO-format bounding boxes, accompanied by random-area train/validation splits and rigorous manual quality control, plus a benchmarking repository with standardized evaluation protocols for semantic segmentation and object detection. This work enables reproducible, comparable research and practical mapping/planning applications, while acknowledging limitations in geographic coverage and class distribution. It also outlines future expansions to include temporal data and additional modalities (e.g., LiDAR) to broaden applicability and robustness.

Abstract

We present NordFKB, a fine-grained benchmark dataset for geospatial AI in Norway, derived from the authoritative, highly accurate, national Felles KartdataBase (FKB). The dataset contains high-resolution orthophotos paired with detailed annotations for 36 semantic classes, including both per-class binary segmentation masks in GeoTIFF format and COCO-style bounding box annotations. Data is collected from seven geographically diverse areas, ensuring variation in climate, topography, and urbanization. Only tiles containing at least one annotated object are included, and training/validation splits are created through random sampling across areas to ensure representative class and context distributions. Human expert review and quality control ensures high annotation accuracy. Alongside the dataset, we release a benchmarking repository with standardized evaluation protocols and tools for semantic segmentation and object detection, enabling reproducible and comparable research. NordFKB provides a robust foundation for advancing AI methods in mapping, land administration, and spatial planning, and paves the way for future expansions in coverage, temporal scope, and data modalities.

Paper Structure

This paper contains 7 sections, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Dataset Sources and Collection
  4. Dataset Structure and Content
  5. Applications in Mapping and Planning Domains
  6. Discussion
  7. Conclusion and Future Work

Figures (2)

  • Figure 1: Examples from two locations in the NordFKB dataset. For each location, the top image shows bounding boxes and the bottom image shows the corresponding segmentation mask.
  • Figure 2: Overview map showing the seven geographic areas included in the NordFKB dataset: Bergen, Kristiansand, Rana, Sandvika, Stavanger, Tromsø, and Verdal. The areas are distributed across Norway, covering a range of climatic, topographic, and urbanization conditions.