Proposal talk:Ground Control Point

From OpenStreetMap Wiki
Jump to navigation Jump to search

Move page to Proposed_features/Ground_Control_Point

@Ivangayton: this looks like an interesting tag proposal. Since it hasn't been used yet, the best place to have this page would be under Proposed_features, so for example Proposed_features/Ground_Control_Point. Would it be ok to move this content over there? --Jeisenbe (talk) 04:46, 13 October 2020 (UTC)

Of course! Sorry @Jeisenbe:, I should have done so in the first place.

Key

Since you raised the idea of ground_countrol_point=* (for whatever reason), it should be better to use that directly, given a GCP is more of a property or ascribed feature on an existing feature. It's not man_made=*. I can see you are trying to develop a surer method for this nice idea starting with an explicit northing=*, but what is uuid=*??? ---- Kovposch (talk) 17:56, 15 October 2020 (UTC)

NASA SpaceApps Apps 2021 with www.osm.to

hey there, we would like to work on GCP for this years NASA Space Apps Challenge 2021 hackathon. They have a drone challenge that shall combine satellite and drone imageries and even further data and we would like to tackle this. We started a small project with www.osm.to that allows QR-codes to link to OSM links with the specific coordinates needed for this particular QR-code and its location. When we started this wiki page was not existing, so it would be cool to coordinate here a bit. We want to avoid to invent something new that already exists now :). So, are GCPs now implemted? And do you have a small example for it? --Hornig (talk) 20:36, 20 September 2021 (UTC)

Hi, this is awesome! Thanks for reaching out @Hornig:! We haven't implemented GCPs in OSM yet, but we've got some examples of GCP collections we're using practically in combination with drone imagery. [Here is an example](https://docs.google.com/spreadsheets/d/1-2utw5D2mr0FW-0GCz0Xn9qRN7a9iGpSVodWWJTQr5g/edit?usp=sharing). We'd love to participate in making this better, and we have a [project working with drones](https://omdtanzania.medium.com/omdtz-launches-fellowship-program-focusing-on-emerging-technology-f99d049d5a19) that could both contribute to and benefit from this. Most of our team is based in Tanzania, perhaps we could arrange a call or video chat? --ivangayton (talk)

hi ivangayton, thank you for your reply :). I checked your google spreadsheet and quickly checked with google maps. Do you need natural GCPs or do you have on ground markers?
And what I would like to know is what the OSM already discussed and prepared for GCP data. Are there already a set of input data including suggestions for tag names? It would be cool to already have some GCPs prepared even as spreadsheet data that we could then easily feed into OSM once the new GCP feature is there.
It would be cool to discuss it here and for further chats we have this Zulip chat. --Hornig (talk) 20:49, 22 September 2021 (UTC)
hi again, I just started the NASA SpaceApps 2021 project page. It would be cool to see some of you joining there :). you find the project and the join-team page here https://2021.spaceappschallenge.org/challenges/statements/drones-and-satellites-for-urban-development/teams/osmto/project --Hornig (talk) 19:39, 27 September 2021 (UTC)

Add the location to the tag?

I'm wondering if this proposal is still active. If it is then I would also propose that the actual location (lon/lat) be used in the tag so that if the node is moved then validators can detect that the location of the node is different from the given ground control point location and flag it as a possible issue. Basically the process would be to find the "exact" location of a point in the world and then add a node (or edit an existing node) that represents that point to add a tag `ground_control_point=<lon>,<lat>`. At this point the mapper may or may not want to move the node if it already existed to the correct location in OSM. The main reason not to move the node into the correct place is that other points on the map may need to move and more data may be necessary to move all of it at once. In most cases the mapper would move the node to the location that matches the newly tagged ground control point. This node would most likely stay at this spot "forever". If it is accidentally moved or moved incorrectly then validation tools can identify the nodes that are located at a different lon lat than their GCP tag suggests. Editors like iD and JOSM could also be tweaked to warn mappers when they move a node that has a GCP, but that would be out of the scope of this proposal.

If adding a lon,lat to a tag is too gross then That's fine - we can simply scrape all of the GCP locations and keep them as in a DB somewhere and just compare thier current locations with what we have on file. What do you think?

Some tools eg OSMCha already warns if a man_made=survey_point is moved. It would be redundant to store this again, not to mention difficult to format and parse (viz DMS, decimal minute, decimal degree).
I have the impression of seeing a "calibrated" tag somewhere. Similar to survey_point:datum_aligned=yes (to its datum used), but for OSM purpose (WGS84)
---- Kovposch (talk) 05:47, 27 October 2021 (UTC)

First Tagging

Hi, I have added the ground_control_tag to a man_made object here https://www.openstreetmap.org/changeset/122712864#map=19/32.22047/-101.51251 . Would you please check it and tell me, if this was correct? I am a first time tagger, so I ask for some help here. :) --Hornig (talk) 21:57, 22 June 2022 (UTC)

Different Types

Hi there, I would like to ask for help and ideas, how to make GCPs distinguishable. I am asking because I know different kinds of GCPs and my goal is to have also QR-coded OSM.to GCPs here. For that I want to understand how we could add more data to GCPs. Let me start with the Standard GCPs: They are 2x2 tiles and typically there are black and white tiles. The connecting corners of the two black tiles is the "center" of the GCPs where the location (lon,lat,alt) is known. So this needs to be a tag that the reference point is one point, and the center and with those 3 fields. Next is the purpose of the GCP. I have seen the tag "callibration_target" with values such as satellite. that I like. maybe it could be used that satellites also means for lower flying things like drones. What I also have seen are 2x2 triangles in red and black. These are typically smaller but same principle with the reference point at the center where the triangles are touching. With the qr-coded OSM.to GCPs, it would be minimum one reference point (as seen here https://sa-2019.s3.amazonaws.com/media/images/18200a00-bdbd-42df-a44c-edd20360198b.max-1000x1000.png ). It is the center of the top-left "bobble". But QR-Codes have 3 of these "bobbles" and all of them would improve the geo-referencing of the maps using these GCPs. So a minimum of one (top left) needs to be filled in, but a maximum of 3 could be used. So a set of 3 lon,lat,alt data should be taggable. How would you/we do that? We currently thought about using black white for the contrast but different colours would be possible.

So, what would you suggest here? Did I miss one? Is meta-data missing that needs to be tagged. --Hornig (talk) 20:28, 24 June 2022 (UTC)