Kaart: Road Geometry Guide

Introduction

All objects on OSM have what is called ‘geometry’ - the shape of the object. It is our goal to make the way and area geometries match what we can see in aerial imagery, to the best of our ability. In this document, we will learn basic road geometry and best practices for mapping.

Prerequisites

It is recommended that you know OSM basics like nodes, ways and polygons as well as how to digitize these features. Other recommended trainings before moving onto road geometry are as follows:

• Intro to Editing: Objects in OSM(link)
• Kaart: Road Tagging Guide
• Editing Software:
  • JOSM Usage(link)
  • ID Editor and Rapid Guide(link)

Skill Instruction

Tools We Utilize

Mapping roads and modifying OSM geometries in general require a variety of tools. First, mappers will need editing software such as Rapid or JOSM. Next, in order to verify our edits, an imagery is required to align to. Be sure to verify any OSM edits with multiple sources, when applicable. Check different imagery layers to find the most up to date imagery for the area you are working in. Country-specific guidelines are also important to consider when digitizing roadways, aligning to imagery, or validating data in OSM. Resources for OSM mapping include:

• Mapbox Satellite
• Esri World Imagery
• Local aerial imagery

Other helpful tools for digitizing roadways include:

• Mapillary Streetside
• Karta View
• Strava Heat Map(useful for verifying foot/vehicle traffic)

Please note, it is not recommended to use "Esri World Imagery (Clarity) Beta" since this imagery has been deprecated since 2/7/2022 and is no longer being updated.

Alignment

Mapping What we can see

It is good practice to only map with imagery that you can see clearly. Avoid using imagery that is obstructed, blurry, pixelated, or unclear when possible. When mapping, sometimes there can be difficulties in identifying road geometries. This may be due to blurry/pixelated imagery, when multiple satellite imageries are stitched together poorly, or visual obstructions like clouds, vegetation, and shadows. In these examples, see how clouds, vegetation, and shadows can obscure roadways.

Shadows obstructing view of the road.	Clouds obstructing imagery of roads.	Tree cover obstructing view of roads.

Using the Correct Amount of Nodes

Excess Nodes

Here is an example of a rounded roadway. For geometries such as this, several nodes will have to be used so that the roadway does not appear angular. However, using too many nodes is unnecessary data. In this example, see the difference between 22 nodes and 11 nodes on the same roadway. Note that even with 11 less nodes, the roadway does not appear to have poor geometry nor sharp angles.

Excess nodes before/after.

In many cases, using fewer nodes can benefit the overall road geometry. For these small corners, 3-5 nodes is typically sufficient. In this example, see the difference between 8 nodes and 3 nodes on the same roadway. Visually, the roadway with fewer nodes looks better.

Excess corner nodes before-after.

Deficient Nodes

Just as too many nodes can be problematic, mapping roadways with too few nodes can lead to poor map data. In this example, we can see a rounded section of roadway that has been generalized as a sharp angle. Aligning this roadway to imagery and adding more nodes significantly improves this feature.

thumb

Sharp Angles

As previously mentioned, digitizing a road with too few nodes can cause angular geometry that does not align to imagery nor accurately represent the roadway. However, there are some cases where roads do not need to have rounded corners. In this example, we can see a sharp angle depicted in imagery. Therefore, we can safely map this roadway with a non-rounded corner.

thumb

Splitting & Combining

There are several instances where roadways should be split. In general, roadways are split to define distinct roads that are connected to a larger road network. Correctly splitting and combining roadways gives other mappers a solid basis for adding tags such as one-way tags, road names, and other crucial details for routing.

Stoparchy

The Stoparchy can be visualized by referring to this flowchart:

When deciding where to stop digitizing a road, where to connect it, or even where to change a road’s classification, consider the “Stoparchy”. This helpful resource is a hierarchical system used to determine where a roadway will end. The most desirable way to end a road is to connect it to another road, but this isn't always possible.

Some examples of good places to end a road inlcude:

• Intersections or forks in the road.
  • Notably, these connections might include one or more classifications of highways.
• Changes in surface type.
  • An example of this would be if a dirt or gravel road connected to a paved road. We can typically see these surface changes in imagery.
• The edge of the task.
  • When working on a continuous roadway this is a good place to end a roadway you are digitizing. Similarly, sections of roads sometimes will not be mapped due to imagery issues.
• Ending a roadway at the last building.
  • This is common in rural areas or along track roads.

Connecting Roadways

Roadways should not be combined if they are different roads. In this example, see how these roadways are left as separate ways in OSM. They share a connecting node, but are split into distinct geometries.

Road Junctions

Typically, roadways are split when they converge, at corners, or where a change in the road name is likely. In this example, see how the roadway was split at an intersection. Utilize the "Color Long Roads" setting in the Kaart Style(verfiy) paint style to identify where roads begin and end more easily. For more information about how to map road convergences, refer to Oneways(link) and Road Associated Features: Signs(link).

thumb

Bridges and Tunnels

OSM ways will be split to accommodate bridges, tunnels, and similar features. In this example, see how the road was split into distinct sections on either side of the bridge. These individual segments were then combined at conjoining nodes for routing/navigation purposes. For more information on what tags to use for bridges, tunnels, etc., refer to Road Associated Features: Tunnels, Bridges, Fords, etc.(link)

Intersections

An intersection in its simplest form is when two ways cross and are joined by a node thus creating an intersection point. Highways that cross on the same level should always be joined with a node. In OSM an intersection is also known as a junction which will be covered in detail below.

Junctions

Road intersections are referred to as junctions in OSM. Different junctions are used for different situations to help the user know what to expect at each junction. The general list of junctions are as follows:

• Simple intersections: when two ways share a common node
• Dual carriageway intersections: when a single or a dual carriageway intersects a dual carriageway thus resulting in two or four common nodes.

• Roundabouts: Roundabouts should be tagged as junction=roundabout. Small roundabouts where the middle can be traversed by vehicles should be tagged as highway=mini_roundabout.
• Circular Junctions: Circular junctions are similar to roundabouts, but traffic inside the circle yields to incoming traffic instead of the other way around. These junctions are tagged as junction=circular.
• Traffic control: Intersections with traffic controlling devices. Intersections with stop signs should be tagged highway=stop. Intersections with traffic signals should be tagged highway=traffic_signals.
• junction=yes: This tag is used for junctions that have other properties like a name=*. These can be added as a tag on a node or as an area around the junction.
• highway=motorway_junction: This tag is used to identify junctions along major highways with junction refs or names where you can exit the highway onto another road. These are commonly used on grade-separated junctions whereas junction=yes is used for at-grade junctions.

For more information about junctions see the junctions wiki page.

Public Transportation

When digitizing or adjusting existing road geometry, we should keep public transportation systems in mind. These routes can exist as separate geometry or can share the geometry of existing ways. It is up to you to determine how to add/modify geometry for public transportation. For more information see the Route Relations(link) doc.

MapWithAI

MapWithAI is the JOSM plugin version of RapID. It is most useful when adding in large amounts of missing data into the map. While greatly reducing the number of features to be added by hand, it still requires human validation as the AI has been known to make mistakes. To learn more about how to add roads with MapWith AI refer to the MapWith AI Plugin for JOSM(link) doc.

Suggested Bookmarks

• Intro to Editing: Objects in OSM
• JOSM Usage
• ID Editor and Rapid Guide
• Kaart: Road Tagging Guide