[GRASS-user] Pansharpening Landsat 8 Imagery

[Michele Toma]

Hi everyone,

I was wondering if anyone had any suggestions or workarounds to the issue that I am having with pansharpening Landsat imagery? My email below has the steps that I took to attempt to pansharpen, but once I create a natural color image and run the pansharpen tool, I end up with distorted colors. Are there any other methods that I should try?

Thanks in advance,

Michele

From: grass-user [mailto:grass-user-bounces at lists.osgeo.org] On Behalf Of Michele Toma
Sent: Wednesday, July 13, 2016 5:39 PM
To: grass-user at lists.osgeo.org
Subject: [GRASS-user] Pansharpening Landsat 8 Imagery

Hi everyone,

I am having difficulty with pansharpening Landsat 8 imagery using this tutorial: http://planet.qgis.org/planet/tag/landsat/. I am able to create a natural color image, but the pansharpening results in images with distorted colors.

Here are the steps that I have followed:


1.       Import all Landsat 8 bands to GRASS 7.0.4

a.       r.in.gdal input=C:\Users\Michele\Desktop\LC81100362016082LGN00\LC81100362016082LGN00_B1.TIF output=LC81100362016082LGN00_B1 -k -e

2.       Converted digital number data to reflectance using i.landsat.toar

a.       i.landsat.toar input=LC81100362016082LGN00_B output=LC81100362016082LGN00_refl metfile=C:\Users\Michele\Desktop\LC81100362016082LGN00\LC81100362016082LGN00_MTL.txt sensor=oli8 date=2016-03-22

3.       Color balanced using i.colors.enhance for bands 4, 3, 2 as RGB respectively

a.       i.colors.enhance red=LC81100362016082LGN00_refl4 at Osaka green=LC81100362016082LGN00_refl3 at Osaka blue=LC81100362016082LGN00_refl2 at Osaka

4.       Displayed RGB map using bands 4, 3, 2 (I get a really great natural color image)

5.       Pansharpened bands 4, 3, 2 using band 8 as the panchromatic band with the IHS method

a.       i.pansharpen red=LC81100362016082LGN00_refl4 at Osaka green=LC81100362016082LGN00_refl3 at Osaka blue=LC81100362016082LGN00_refl2 at Osaka pan=LC81100362016082LGN00_refl8 at Osaka output=LC81100362016082LGN00_ihs method=ihs

6.       Displayed RGB map with the "pansharpened" bands for 4, 3, 2

7.       Result: An image that is entirely gray (I found this link that says to use i.fusion.hpf, but following those instructions, I get "'make' is not recognized as an internal or external command, operable program or batch file." http://gis.stackexchange.com/questions/111763/pan-sharpen-landsat-8-using-grass-7 Am I supposed to use the GUI for this?)

>From searching online, I have found that the i.pansharpen module doesn't take 16 bit imagery (which Landsat imagery is) but rather 8 bit. I found this link (http://gis.stackexchange.com/questions/39174/how-to-color-balance-false-composites-of-pan-sharpened-landsat-images-in-grass/67319#67319) that says to rescale the image. I attempted this twice:


1.       When I run it on the bands produced after running the i.landsat.toar module, I get numbers such as min=-0.00813278402972343 and max=1.58301384751301. After rescaling the images to 0-255 and proceeding with step 3 I end up with an entirely white image.

2.       However, if I rescale the original bands which all have ranges of 0-65535, to 0-255, then pansharpen, the result is a washed out version of the natural color image. I believe this has been pansharpened, but the colors are off. I also attempted to color balance the pansharpened bands but was unsuccessful as well.

Am I rescaling the images incorrectly or is there another step I am getting wrong? Any guidance would be greatly appreciated.

Thank you,

Michele Toma
michele.toma at airborne.aero<mailto:michele.toma at airborne.aero>
www.airborne.aero

[New Small2]

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20160718/e0fc63da/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 2598 bytes
Desc: image001.jpg
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20160718/e0fc63da/attachment.jpg>