7 replies to this topic

[MikeH]

I downloaded a .bil file from the USGS Seamless Server and I am attempting to color it using Photoshop's Gradient Map feature. I have the Photoshop end nailed down*, but my problem seems to be with the original .bil file. It seems to have at least three grayscale bands at different elevations for this particular region. (Regions with greater relief have more bands). I think I understand what's happening is the grayscale starts at the lowest elevation in black, then goes to white at about 500' or so then starts over again with another black to white band.

My question is: How can I get data with only one grayscale for the entire region, and/or, how can I get Photoshop to draw the area with one grayscale?

This map is for illustration purposes only and does not need to be scientifically accurate in elevation. It just needs to look good.

*I'm opening the .bil as a .raw file in Photoshop, and using the information in the .hdr file to define the pixel width x height. I then use Photoshop's lighting effects tool on another layer to render the 'bumps' and map a color gradient ( blue-sea level to green to yellow to brown-high areas) to the original background layer. The end result looks great, except for the fact that I have grass growing on hilltops that should be brown.

Here is an example. Not bad for my first attempt at a shaded relief.

 Shaded_Relief_Test.jpg   400.22K   106 downloads

[frax]

What do you mean by three bands? Are there there channels? Something sounds a bit odd here... Is it 16-bit or 8-but? I assume 16-bit... ?

[MikeH]

Yes, Photoshop is opening the original .bil as a 16 bit .raw image, but I have to convert it to 8 bit in order to run the lighting effects and color gradient. The example I posted shows colors going from green to brown, as the elevation gets higher going up from Folsom Lake (CA, USA), then shifts back to green at a certain elevation. And again at the next higher elevation.

So basically, the brown and green area east south east of the lake is actually higher than the green area in the south west corner of the image, but the origional .bil file has the same grayscale in both areas which causes Photoshop to map the same colors over again.

Does this help clear things up?

[MikeH]

Here is the original grayscale image from another region (southern California, LA area) that shows the banding. It's most evident in the mountain regions. This is exactly what I see when I open the .bil file in Photoshop as a .raw.


Oh, and Photoshop is only showing me one channel - grayscale.

 Relief_Map.jpg   886.31K   94 downloads

[frax]

I have never seen this - I think there is something funny going on with the gradient that is applied. Are you sure that there is no color map or something? Maybe it is related to the import of the raw image, you can play with the parameters there.

[MikeH]

Unfortunately, Photoshop doesn't give me many options for importing the RAW data file other than pixel width and height. I realize that Photoshop is not the preferred tool for this kind of job, but it's all I have access to at the moment.

My main problem can be seen two different ways:
1) The data I'm getting from the USGS seamless server has too many elevation bands of grayscale (maybe too much detail for what I'm trying to do.)
or
2) Photoshop can't do anything with elevation data so it maps color according to grayscale only.

Since I can't do anything about #2, I was hoping I can download a digital elevation map that uses one grayscale for all elevations. Does anyone know where/how I can do this?

[frax]

Can you look at the .bil in another software or a hex editor? Is there a separate color map? I am pretty sure that there is no grayscale color bands stored with the image. I think you can get to some settings in Photoshop where you specify the interleave and endianness for raw images, can't you?

If you can put the image up for download, I could take a peek at it.

[mdsumner]

It looks exactly like a number range interpretation issue as mentioned. When you say "band" above I think you are meaning "range of values" within an attribute, such as negative numbers are marine, positive are land and so on.

I would check the metadata with gdalinfo filename -stats and perhaps use gdal_translate to convert to something useable by other software.