Jump to content

 
Photo

Hillshade Problem

- - - - -

  • Please log in to reply
5 replies to this topic

#1
burwelbo

burwelbo

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 113 posts
  • Interests:Hockey, Hiking, Travel
  • Canada

I have been working on a hillshaded relief map for an area in Canada and came across a resolution problem within the data. It looks like the data has been merged together from different resolution datasets and the difference is very noticeable. The data was downloaded from the GeoBase web site. I have attached a sample. Does anyone have any suggestions how to fix it? I am using ERMapper, ArcMap and Photoshop.

Thanks in advance
Bruce

Attached Files



#2
Charles Syrett

Charles Syrett

    Ultimate Contributor

  • Validated Member
  • PipPipPipPipPipPip
  • 529 posts
  • Canada

To state the problem in another way -- in some areas there's more topographic detail than in others, because the grid (or DEM) was created from sources that varied in resolution. The only way I can think of (since you can't upsample, you can only downsample) is to guess where the more detailed section of the DEM grid is, downsample it, and then re-merge with the other section. This would be a very hairy, time-consuming, and trial-and-error process. And what would be the point? Most people will never notice the resolution difference on your hillshade -- in fact I found it hard to notice myself. Will you be adding more map detail over this image? If so, I would definitely not spend any time fiddling with it.

Charles Syrett
Map Graphics
http://www.mapgraphics.com

I have been working on a hillshaded relief map for an area in Canada and came across a resolution problem within the data. It looks like the data has been merged together from different resolution datasets and the difference is very noticeable. The data was downloaded from the GeoBase web site. I have attached a sample. Does anyone have any suggestions how to fix it? I am using ERMapper, ArcMap and Photoshop.

Thanks in advance
Bruce



#3
Pete

Pete

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 151 posts
  • Gender:Male
  • Location:Inverness
  • United Kingdom

I see what you mean. Charles is right - it's not like you can add more information to your DEM without having the raw data to work with and manually altering the data would take ages.

The only thing I can think that you might want to try is to use Photoshop to make the cuts in the data less obvious. For example, in the top right of your image where the resolution of the data noticably changes resolution along a ridge line, you could use Photoshop to duplicate your shaded relief and then blur the duplicated layer to the point where it looks similar to the lower resolution, then simply use a layer mask and cartographically Tipp-Ex over the sharp detail with the blurred layer.

I used this technique to touch up the top right corner and it works alright:

Attached File  blur_animation.gif   170.16KB   105 downloads Attached File  blur_animation.gif   83.47KB   112 downloads

The second animation flips every 5 seconds so you can see the difference better.

I would only use this in areas where the contrast in resolution is partculary bad to soften the transition, I certainly wouldn't use it to selectively replace whole parts of the DEM simply because it would take too long. You can see from the animation that the effect is very subtle - is it worth it ;) ?

It's a really dumb question but do you have other things you are going to put on the map? I'm only asking because I've managed to hide holes in my data by covering them up with a title, text, scale bar or something like that. It might not look nearly so bad when the map has more on it.

#4
burwelbo

burwelbo

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 113 posts
  • Interests:Hockey, Hiking, Travel
  • Canada

There is going to be more map data on top. MAybe I can hide much of the problems that way. I will try using photoshop blur the transition to make it look less noticeable.

Thanks
Bruce

#5
graveldraw

graveldraw

    Contributor

  • Validated Member
  • PipPip
  • 23 posts
  • Gender:Male
  • United States

I have been working on a hillshaded relief map for an area in Canada and came across a resolution problem within the data. It looks like the data has been merged together from different resolution datasets and the difference is very noticeable. The data was downloaded from the GeoBase web site. I have attached a sample. Does anyone have any suggestions how to fix it? I am using ERMapper, ArcMap and Photoshop.

Thanks in advance
Bruce


Probably just my eyes, but I don't see a block within the image that appears to have a different resolution. In my work as a soil scientist, I see differences between mappers, topographic maps, detail of streams, and DEM: but not here. I'm wondering where this area is? I do see the differences in the degree of dissection of the ephermal drainageways (corrugations) along some of the flanks of the mountains or ridges. These corrugations might be due to differences in geology and thickness of colluvium rather than DEM sources.

I agree that this DEM has a little too much detail for your map. Resampling or blur would probably be the way to go.

Dan

#6
Clark Geomatics

Clark Geomatics

    Jeff Clark

  • Validated Member
  • PipPipPipPip
  • 105 posts
  • Gender:Male
  • Location:North Vancouver, BC
  • Canada

I have been working on a hillshaded relief map for an area in Canada and came across a resolution problem within the data. It looks like the data has been merged together from different resolution datasets and the difference is very noticeable. The data was downloaded from the GeoBase web site. I have attached a sample. Does anyone have any suggestions how to fix it? I am using ERMapper, ArcMap and Photoshop.

Thanks in advance
Bruce

Have you tried merging your lower resolution areas with SRTM data? Although the SRTM data is of lower resolution than the CDED data you are using, it might improve the look somewhat. Just thinking out loud...
Cheers,

Jeff Clark
Principal
www.clarkgeomatics.ca




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->