StreetSurfaceVis: a dataset of crowdsourced street-level imagery annotated by road surface type and quality

TL;DR

This paper introduces StreetSurfaceVis, a large annotated dataset of street-level images for road surface assessment, and proposes strategies to balance class distribution and reduce manual annotation effort.

Contribution

We present a novel, diverse dataset of street images annotated by surface type and quality, along with a sampling strategy leveraging external resources to improve annotation efficiency.

Findings

The dataset contains 9,122 images from Germany.

Our sampling strategy reduces manual annotation workload.

Class imbalance is effectively mitigated using external label prediction resources.

Abstract

Road unevenness significantly impacts the safety and comfort of traffic participants, especially vulnerable groups such as cyclists and wheelchair users. To train models for comprehensive road surface assessments, we introduce StreetSurfaceVis, a novel dataset comprising 9,122 street-level images mostly from Germany collected from a crowdsourcing platform and manually annotated by road surface type and quality. By crafting a heterogeneous dataset, we aim to enable robust models that maintain high accuracy across diverse image sources. As the frequency distribution of road surface types and qualities is highly imbalanced, we propose a sampling strategy incorporating various external label prediction resources to ensure sufficient images per class while reducing manual annotation. More precisely, we estimate the impact of (1) enriching the image data with OpenStreetMap tags, (2) iterative training and application of a custom surface type classification model, (3) amplifying underrepresented classes through prompt-based classification with GPT-4o and (4) similarity search using image embeddings. Combining these strategies effectively reduces manual annotation workload while ensuring sufficient class representation.