HERE map formats can be stored and exported in various data formats, depending on the map proudct and the needs of particular customers. Regardless of format, this map data can be searched and filtered in powerful ways to facilitate common use cases. Similarly, the quality of this data can be assessed and rated by software validation rules and ground truth. For details, see the sections below.
HERE provides map data in as many as three formats, depending on the product: our Native format based on Protocol Buffer, the Navigation Data Standard (NDS) format defined by the NDS Association, and the Map Object Model (MOM).
Searches of HERE map data can be filtered by feature (e.g., topology segments), location (map tile level and ID), attribute (e.g., road signs), sub-attribute (e.g., stop signs), and quality Spec Level (e.g., level 4 as explained below). For example, you can filter HD Live Map data in a Level 14 tile in San Francisco for stop signs meeting or exceeding the quality and requirements of Spec Level 4.
The output of filters includes highlighting positive search results in a map and listing those results in a table.
The Quality Index is a numerical measurement of the confidence HERE has in the correctness of specific map data. Accuracy (absolute and relative) are two of the explicit measurements, among many, composing the Quality Index.
A collection of confidence models based on the Quality Index, and a few other factors, are used to determine the Confidence Indicator for a given attribute or feature of our maps. We share Confidence Indicators with our customers. For a HERE map product to be approved for production and made available to customers, all the features and attributes in that product must meet or exceed the target Confidence Indicators, relative to ground truth.
For example, the Barrier Classification attribute assigns a type to each and every barrier in the map. Currently supported types of barriers include Guardrail, Jersey Barrier, and Flat Wall. The quality measurement that makes sense in this case is "Error Classification Rate" which measures the percentage of barriers in the map that are misclassified relative to reality.
Map data can be filtered based on the type of data and the values of Confidence Indicators.