OpenStreetMap
Gregory Peony's Comments
| Changeset | When | Comment |
|---|---|---|
| 158316372 | about 1 year ago | Most of the building footprints here are inaccurate. Again it's a question of imagery interpretation. Features in imagery; like trees, walls or buildings can obscure others. In situations like these you should try to work out what is the shape of a building (in this case) likely to be. Sometimes different imagery sources can help with this, other times you'll have to give it your best guess. In this case all of the building footprints should have been squared, given the low resolution imagery you were working with and the architecture of the surrounding buildings. It was more likely that the imagery was distorted or that the building was partially obscured by trees/vegetation than it is for it to have this unique footprint. Buildings tend to take take on this shape near highway intersections of 'odd' angles. I hope this helps.
|
| 158315601 | about 1 year ago | Overall this is a high quality contribution. Well done! All buildings digitized in this changeset are in fact buildings, and the majority of them have accurate footprints, and are squared appropriately. There are some inaccurately shaped outliers however, which I believe is due to you missinterpreting the imagery. I encourage you use Mapbox imagery as a suplimental source, because it has a greater resolution in this area. That should help you better determine the shapes of buildings. If the available imagery is if poor quality you should interpret it and add buildings with likely shapes. Ask yourself: What shape is that building likely to be? Is it actually that shape, or does it just look that way on the imagery? Thank you for your contribution!
|
| 158254555 | about 1 year ago | Buildings and roads should rarely overlap or share nodes. That may be the case at a petrol station where the road passes under a 'roof' but generally they should not becasuse it could affect some routing software. When in close proximity to other digitized features hold Alt to prevent your cursor from snapping to them. I have modified these features to provide an example. Some of the building 'outlines' do not match up with the roofs in the imagery. This is because we are aiming to map the building footprint (where the walls meet the ground) so that the map is accurate in 2D. Check the OSM carto imagery layer to get an idea of what your mappings look like to others. Thank you for your contribution. I hope my feedback helps you to improve the quality of your mapping.
|
| 158255095 | about 1 year ago | Well done! You identified buildings in the imagery and I don't think that you outlined anything other than a buiilding. The buildings' footprints however, are not accurate. The low resolution imagery does not help, but you need to interpret the imagery and give buildings realistic shapes rather than tracing distorted imagery or artefacts. I recommend that you find a task area where Mapbox imagery shows on the whole the same buildings, because it has a greater resolution in this area. This will allow you to compare what the same buildings look like with different levels of resolution. When you must work with low resolution imagery zoom out first to ascertain the shape and orientation of buildings, then zoom in to digitize them accurately (you can repeat this of course if you forget the shape for example). Quality > Quantity. Thank you for your contribution.
|
| 158206496 | about 1 year ago | Please square your footprints if a building's walls form right angles. The feature I flagged consists of a surplus node. Overall this is a good contribution; all features are valid. Zooming out can help you see the shape and orientation of features, then zoom in to draw the feature accurately. If you want to you can use mapbox imagery in this area (either to trace or as a supporting reference) by offseting it to allign with the features you've already mapped, since the resolution is greater. I encourage you to attend this online event if you'd like to interact with other mappers. https://osmcal.org/event/2725/ or search for it through OSMCal
|
| 158290227 | about 1 year ago | I made some modifications to the buildings in this changeset, hopefully for you to use as a reference. Thank you for your contribution.
|
| 158290227 | about 1 year ago | Well done! You have correctly identified all of the buildings digitized in this changeset, and the vast majority of footprints you've drawn here are quite accurate, although they should be squared (q on keyboard after tagging). You've done a good job placing the nodes with your cursor, but it's very unlikely that you'll get it spot on, and the q keq will make the corners right angles. The feature I flagged has too many nodes, and I think it's actually two buildings given that the shadow is cast West in the imagery you used to trace it and there is a dark line between the two light parts.
|
| 157085950 | about 1 year ago | Although I have verified this changeset as bad, there is a lot that I like about it too. For instance you've squared the majority of rectilinear buildings that you've added in this changeset. In places you followed good practice and modified features rather than deleting the old ones and adding new ones in their stead, (and in some cases improved existing mapping). I also like that you decided (seemingly part way through) to use an alternate imagery source in this area, because it had a greater resolution. If you want to improve the first points I'll give you is to draw building footprints slightly smaller, ~95% the size the roof you see in the imagery. Hold the Alt key when adding nodes close to existing features to prevent your cursor from snapping to them.
|
| 157085950 | about 1 year ago | Although I have verified this changeset as bad, there is a lot that I like about it too. For instance you've squared the majority of rectilinear buildings that you've added in this changeset. In places you followed good practice and modified features rather than deleting the old ones and adding new ones in their stead, (and in some cases improved existing mapping). I also like that you decided (seemingly part way through) to use an alternate imagery source in this area, because it had a greater resolution. If you want to improve the first points I'll give you is to draw building footprints slightly smaller, ~95% the size the roof you see in the imagery. Hold the Alt key when adding nodes close to existing features to prevent your cursor from snapping to them.
|
| 156886825 | about 1 year ago | Thank you for your contribution. You correctly identified the vast majority if not all of the buildings that you digitized in this changeset. Their shape however is not very accurate and a number of features overlap and share nodes. Refer to the editing shortcuts in the help section of the ID editor for 'squaring' and 'avoiding cursor snapping'.
|
| 156886751 | about 1 year ago | Well done, all of the features that you added are in fact buildings. Their shape however, is generally not accurate; all of these building footprints are actually recti-linear, and so should have square corners (press q on the keyboard after accurately tracing). To be fair the imagery you used to digitize the footprints does not have the greatest resolution. If you choose to contine contributing I recommend that you find a project with clearer imagery to gain familiarity with building shapes.
|
| 151799532 | over 1 year ago | WAY: 1286509577 is a T shaped building, again because of the illumination and shadows on the roof.
|
| 151799532 | over 1 year ago | WAY: 1286509640, WAY: 1286509615 (,and others) are L shaped buildings, look at the way the roof is illuminated and where the shadows are to help you determine the shape of buildings. N.B. the shadows in this imagery are cast to the SW and secions of the roof facing the sun will appear brighter.
|
| 151799532 | over 1 year ago | I do not believe that WAY: 1286509575 is a building, because it does not cast a shadow similar to other nearby buildings.
|
| 151855368 | over 1 year ago | The white part to the East of some buildings here appears to be the buildings' wall. This should initially be ignored when using the roof to map a footprint. Then the footprint should be moved to align with where the building's wall meets the ground. Otherwise the footprint will be oversized and misalligned with nearby features.
|
| 151796419 | over 1 year ago | Good job extending the footprint of the Western buidling which is partially obscured by vegetation.
|
| 151796419 | over 1 year ago | The building footprints you drew here are too large. A building's footprint is usually <= the roof shape/size. I modified these footprints and you can see that via their history.
|
| 154149821 | over 1 year ago | oh, and you're welcome!
|
| 154149821 | over 1 year ago | You can use this as an aid when determining the classification of highways https://www.canva.com/design/DAFrDF4ptto/AjjrER_Dd7d-gjSh6jJdgw/view?mode=prototype#main-page , source: UN Mappers Published under Creative Commons Attribution – ShareAlike 2.0 Generic
|
| 154149821 | over 1 year ago | I think that the majority of highways tagged as residential e.g. WAY: 1302540340 in this case should instead be unclassified, becuase these highways serve to connect different settlements via the highway network or they go through the settlements, and the largest settlements connected by them are villages.
|