High-Definition Map Cross-Layer Spatial Consistency Benchmark

#spatial reasoning #cross-layer consistency verification #geometric relationship analysis #high-definition map validation #autonomous driving navigation #road infrastructure mapping

Records: 5000
Size: 0.4GB
Format: JSON
Provider: MOBIUSI INC
Updated: v1.02026-11-04

AI Analysis & Value Prop

Designed for evaluating urban road HD maps, this benchmark covers GIS layers such as lane markings, lane centerlines, road boundaries, and road facilities, together with feature geometries and coordinate reference information. It provides spatial relation queries and reference judgments for positions, alignment errors, and containment across layers or features, enabling assessment of a model's understanding of spatial relationships in multilayer HD maps. Typical applications include 2D spatial understanding evaluation, map data quality checks, and related model development.

Overview

Dataset Insights

Preview

Sample Examples

sample.json

{}

Schema

Technical Specifications

Field	Type	Description
file_name	string	File name
gis_layer_data	object	Raw GIS data for layers such as lane markings, lane centerlines, road boundaries, and road facilities, including features, geometric coordinates, and coordinate reference information.
relation_query	string	Describes the two layers or features being evaluated and the spatial relation to determine.
relation_exists	boolean	Indicates whether the spatial relation stated in the query holds.
alignment_error_m	number	The alignment error in meters, calculated from the geometric positions of the relevant layers or features.
containment_status	string	Records the containment judgment between target features; not applicable indicates that the query does not involve containment.
reference_relation	string	The spatial relation category established from layer geometries and reference annotations.
reference_explanation	string	Explains the geometric position, boundary alignment, or containment evidence supporting the reference spatial relation judgment.

Trust

Compliance Statement

Authorization Type

Proprietary - Commercial AI Training License (No Redistribution)

Commercial Use

Requires exclusive subscription or authorization contract (monthly or per-invocation charging)

Privacy and Anonymization

No PII, no real company names, simulated scenarios follow industry standards

Compliance System

Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

FAQ

Frequently Asked Questions

What map content does the benchmark cover?: The data covers HD map GIS layers such as lane markings, lane centerlines, road boundaries, and road facilities, with related features, geometries, and coordinate reference information. The exact sample size and geographic coverage are specified on the product page and in the delivery documentation.

What spatial relation annotations are provided?: Each record includes a spatial relation query and a reference relation category. It may also provide a relation-existence label, alignment error in meters, a containment judgment, and a reference explanation to support checks of position, alignment, and containment.

How can I obtain and use the benchmark?: Use the access channel provided on the product page to review the data package and delivery documentation. You can parse the GIS layers and queries using the field descriptions, then compare model predictions with the reference relations and geometric judgments.

What are the licensing terms and intended use cases?: The benchmark is suitable for domain evaluation of HD map spatial understanding models, map data quality checks, and related algorithm development. Check the product page or license agreement for permitted uses, commercial rights, and redistribution restrictions.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

High-quality data, ready for your next AI breakthrough.

Cite this Work

@dataset{Mobiusi2026,
  title={High-Definition Map Cross-Layer Spatial Consistency Benchmark},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/3541608b644bc206a5d9dd006b6e74eb?dataset_scene_id=1},
  urldate={2026-10-10},
  keywords={HD map benchmark, GIS spatial relations, cross-layer consistency, 2D spatial understanding, lane marking data, road boundary data, map quality evaluation, spatial reasoning benchmark},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task