Mapillary images are a good source to verify and add information like turn restrictions and speed limits. This is not possible with satellite images. We realised that we need to have a thorough knowledge about the traffic signs used in the region we are mapping, as they are region specific.
Initially we started exploring the traffic data in USA. With help of the traffic data wiki for USA, we came up with a list of road signs which can be mapped using the Mapillary imagery. This list was created after pruning a list containing 467 traffic signs used in USA, after comparing them to the signs detected by Mapillary. We have also listed the OSM tags associated with each traffic sign.
Refer to the table given below to match the traffic sign in the Mapillary image against the list of traffic signs and it also consists of the OSM tags:
Here is the traffic signal guide.
This is the workflow we used to map the turn restrictions:
STEP 1: Setting up JOSM for mapping traffic data
In order to be able to map traffic data easily, there are some plugins that JOSM needs to be configured with:
- Layers to be added:
- Mapillary layer:
- Bing imagery:
- Mapillary layer:
- Plugins to be downloaded:
- Mapillary plugin for JOSM: This plugin allows the user to view mapillary images in JOSM.
- Turn restrictions plugin for JOSM: This plugin allows to add turn restrictions to the selected roads with ease.
- RoadSigns plugin for JOSM: This plugin offers a dialog for selection of traffic signs from a list of presets. With simple clicks you can create combinations of signs and generate tagging suggestions.
- Map-paint styles to be added:
- traffic_signs: It displays all the existing traffic signs that have been already mapped on OSM.
- Maxspeed: It displays the already existing speed restrictions on OSM.
STEP 2: How to use the mapillary and turn restriction plugins
Using the Mapillary plugin
- The Mapillary imagery can be added by clicking on
imagery -> Mapillaryin JOSM. This will add the mapillary layer over the dataset layer.
The explanation for how the plugin works can be found here.
Note: - If one is using
tasking manager, first download the task from the tasking manager and then add the
- Every time a new task is downloaded from the tasking manager, a new
Mapillarylayer must be added after deleting the old one.
Using the turn restriction plugin - The explanation for how the turn restriction plugin works can be found here.
STEP 3: How to add the traffic data using mapillary
- Probe the different Mapillary images present in that area to find useful traffic data that can be added to the Open Street Map.
- The Mapillary point on the GPX trail that has a
red trianglealong with it contain
Probe the images to get a good idea of the surroundings and then compare them with the existing imagery to find the exact location of the
To add a turn restriction using the turn-restriction plugin
- We need to cautious while adding a
U-turn/No U-turnin OSM, here is how we can add it:
- To get a better idea about the area/road, always inspect multiple images.
- With the help of the plugins, look for one-ways, speed-limits and traffic restrictions which are already present on the road/area.
- Unless you are sure about the start and end point of a road, it is advisable to avoid adding
oneway=yestag to it.
We should avoid adding
turn restrictionto roads that are leading to oneways going in the opposite direction, like the example below. It will only result in redundant data:
Here we should avoid adding
turn-restriction as it will result in redundant data
While mapping, we came across the following issues:
- They are easily visible in the imagery and map but are leading to
splittingof major roads and also leading to addition of multiple
- Is this something we need to follow while adding the turn restrictions or there is a better-way?
- It is difficult to identify the start and end point of the road to which the speed-limit should be applied.
- Is there a better way to add speed limits or any convention that is followed while adding them?
For further details about how we created the workflow please refer to this repository.