OpenStreetMap
Gregory Peony's Comments
| Changeset | When | Comment |
|---|---|---|
| 170695959 | Welcome to OSM. You correctly identified a building in the imagery, but mapped its footprint inaccurately becasue you did not account for the fact that a part of its roof is obscured by vegetation. See how I mapped it here https://osmcha.org/changesets/170696131 Please keep this feedback in mind when contributing in future. Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/
|
|
| 170406548 | Hi Annabel, well done, it's clear to me that you are begining to grasp the concept of correcting a footpritn's position when mapping using imagery captured at an angle. There are a few inaccuracies in this changeset I have mapped these footprints as an example in https://osmcha.org/changesets/170626506. --- Before mapping take a moment to (litterally) note two fundamental properties of any imagery you're looking at; 1. The direction from which the scene is lit or shadows are cast. 2. The angle at which the imagery was taken i.e. which sides of buildings you can see. It appears to me that you did not account for the off-nadir imagery when buildings were close together and obscuring one another's walls. In cases like this either another imagery source which was captured from a different angle should be used, or adjecent groups of buildings where the base of at least one of them is visible should be selected (shift+LClick) before accounting for off-nadir and moved (m) in usison to position them more accurately (this works best if the height of the builings is similar). In OSMCha click the map icon and select ESRI which is closer to a vertical photo and see how some footprints are positioned well with respect to it while others are inaccurate. Please keep this feedback in mind when contributing in future. Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/ |
|
| 134375807 | Thank you for the context of the place import. I don't know if a tool already exists, but I agree that it should be possible to do it in a more automated way. I used the balloon tool to create the geometry and replaced an old way per closed loop of the relation to preserve history. Then deleted the excess ways and the type tag needs to be deleted. All of which I think could be semi-automated. Modifying relations tends to be much easier in JOSM. I don't expect you to go back and change them all or anything like that (especially in ID). It should make it easier to map using simple areas in your future contributions. |
|
| 170380357 | You're welcome! |
|
| 170431193 | I have a question regarding the removal of GNS tags. Do they not serve any purpose, and so is it valid to remove them in general or are there conditions for when to do it? It's just a hunch of mine but, I thought that maybe they could prevent future imports from creating duplicate data, so thus far I have not been removing them; maybe you know something I don't. Then again, the place names and source date tags probably should suffice to allow for prevention of duplication.
|
|
| 169397600 | You're welcome! Thanks for fixing the issue, and the info regarding GNS and location accuracy. |
|
| 169397600 | Hi, I'm just pointing out that I spotted an Ikpai 1.5 km SW that's been there for years. Seems like you added a duplicate to me. What do you think? Maybe you had more info, otherwise I recommend using the OSM standard layer to see place names or downloading places in the nearby area to preserve their history. I am not confident about the location of the older POI and the settlement between the two seems to have no name currently.
|
|
| 169397600 | Hi, I'm just pointing out that I spotted an Ikpai 1.5 km SW that's been there for years. Seems like you added a duplicate to me. What do you think? Maybe you had more info, otherwise I recommend using the OSM standard layer to see place names or downloading places in the nearby area to preserve their history. I am not confident about the location of the older POI and the settlement between the two seems to have no name currently.
|
|
| 134375807 | See https://osmcha.org/changesets/170413012 for my resolution of the residential areas mapped using multipolygons. While we're chatting, do you know to make Ovum the alt name of Owom? Just because they're similar words?
|
|
| 134375807 | Thanks for your informative and swift response. I was asking to see if I had missed some other reason as to why it would be better to map it the way you had. It great to have confirmation. I forgive you and accept your appology ;) Now I'll tell you about what I presume to be a fairly unknown feature in the ID editor, to make mapping similar features with simple areas that share nodes much more convenient. The two buttons you need to know are F and backspace. This works with lines and areas. Say we were to re-map the residential areas here using parts of the forest geometry. Select area (3) and click on a node you want to be shared by the residential area and forest, then click another in the direction you want to travel along the existing area. Now press or hold F to add shared nodes and use backspace to remove them. You can also click to add (non shared) nodes freely and follow multiple existing lines or areas. No need for splitting or messing around with relations. I hope this helps and is easier than the methods you are currently using. For a full list of ID shortcuts press "?" They aren't all discussed in the help section, so you may want to experiment and see what's possible. |
|
| 134375807 | Hi, the changeset comment here is invalid: it does not relate to any feautre you edited. The relations themselves are valid however, provided the forests were mapped similarly to how they are now; it seems unnecessary to map these residential areas as multipolygons. They are not enveloped by another relation in which they should be inners, and so could just simply be mapped as residential areas. Mapping them this way means that other contributors will just have a harder time editing them. If you remember; why did you map the residential areas in this way? See osm.wiki/Relation:multipolygon#Mapping_style,_best_practice for a comparison of two methods of mapping multipolygons. Please keep this feedback in mind when contributing in future. Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/
|
|
| 170403318 | Most of the footpritns you added are valid and all have been appropriately squared, but some could be more accurate and others envelope multiple buildings. good job updating/improving some footprints here by modifying the exisitng data instead of deleting and re mapping. Zooming in and out can help a lot to idnetify features and se their shape and orientation. See how I mapped some of the buildings here https://osmcha.org/changesets/170405346 Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/
|
|
| 170402539 | All footpritns you added here are valid and quite accurate. They could be slichtly smaller. Generally pitched roofs overhang the walls of a building, so a footprint slightly smaller than the roof is accurate. You can scale selected features in ID with shift+(-/+), or JOSM with ctrl.+alt+Lclick & drag. Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/
|
|
| 170380357 | The offset may indeed be different from task to task especially if the area is hilly/mountainous. (depending on the context) For more consistency across larger areas it can be beneficial to align the highways which will run through other tasks so that a consistent alignment can be achieved. This works well when the area is quite flat. Remember to add the offset to the comments as the first thing you do after locking a task for mapping; I've certainly forgoten to state which sources I used and their offsets on a few occasions. It's easy to forget if everything already looks aligned. You can use the arrow keys at the sides of the offset box for fine adjustment. In JOSM if you have any questions pressing F1 for context help tends to work well, otherwise check JOSM forums/chanels, or ask. In order to replace geometry you need to install the "utilsplugin2" plugin via preferences (F12) then plugins, search for it and download it. you can click the more info link to see what other functions it enables. You need to replace before you upload by selecting both features and using the shortcut; it doesn't work on features already on the OSM servers. One thing I forgot to mention is that the situation in this area has changed significantly so only a few features may be common between ESRI and Bing, but it can still be useful. |
|
| 170380357 | When using multiple imagery sources you should aim to align them at ground level which makes them easier to work with, features like fences and walls tend to be the best references for this; they have clear edges and are close to the ground. Regarding project 12803 you are requested to offset the imagery to align with map data becasue it's a long lived project and the imagery has changed. When this happens the position of features may move because their position is claimed to be within X m of the true location.
Though few contributors do it, adding imagery offsets you used in tasking manager comments is highly recommended. It's beneficial for subsequent contributors becasue they can be coppied, and it also shows if you're correctly offsetting imagery. The best way to do this is to copy and paste the offset you used into the comment box as soon as you've determined it. There is no need to add the offset to changeset comments becasue when the imagery changes it likely won't be valid anymore. I figure it likely wouldn't cause any harm, but so far I haven't seen anyone add offsets to changeset comments. I stated a few times which source was the one I was aligning to but currently I jsut let the order of the sources imply that information: in JOSM I tend to have my primary source in layer 1. |
|
| 170380357 | Hi, most of the footprints you added here are valid. I flagged some footprints here which I think are clearly inaccurate or envelope multiple buildings. The imagery resolution for the primary source here is poor, so think about the most likely shapes of buildings; you are liable to be mislead by the imagery otherwise. Zooming out then back in IMO is crutial to better seeing identifying a building, its shape, and orientation. I highly recommend referencing a secondary source probably bing if you're using ID. I have posted about the imagery metadata, capture dates and offsets to use in the project comments. In this case the bing offset would be 2.45; 1.69 . Do you know how to offset imagery?
|
|
| 170366283 | Are you using the drop down menu of past comments, when you're applying comments to your changesets?
|
|
| 170366283 | Hi, it seems to me that you tend to use very generic and poor changeset comments whenever you map and save changes to multiple features simultaneously i.e. in one changeset. Be deliberate with your uploads/saves. Work on features that have a logical and/or geographic theme between each save: modifying the name of a fast food resturaunt has virtually nothing to do with transportation and power infrastructure. In this case I probably would have mapped this using 4 changesets (each with a good comment) as follows; 1. Change the name of a resturaunt.
Doing this will make it easier for others to review your edits, work with them, and provide feedback and comments. In case you forget what you worked on, the changes you made will be displayed in the panel that opens on the left hand side, when you save your work.
|
|
| 170365762 | You correctly identified and tagged a bridge in the SW of this changeset. Thank you for your contribution. If you want to experience the OSM community or to get timely feedback from other mappers; I recommend that you attend a mapathon. You can find events here https://osmcal.org/
|
|
| 170365696 | In this changeset some of the power towers are positioned slightly North and West of where they should be. Aim for the centre point between the supports (legs). I know the imagery resolution here is not amazing, but looking at the shadows should help you more acurately position the towers. The shadows are cast NW in this scene and the point your aiming for should be on the same axis (line) as the shadow(s). There are some imagery distortions here, so I recommend using the bottom half of the visible shadow to obtain the axis. Next (if possible) try to identify where the supports meet the ground. You may find that pressing "w" to toggle wireframe mode may help you see the underlying imagery. You can move selected points by pressing "m" to move it with your cursor, or with "shift+arrow key". The ammount of shift depends on your zoom level. try them out and see which you prefer. For a full list of ID editor shortcuts press "?". I hope this helps.
|