RGB and CIR visualization

[Jaime Lopez Carvajal]

Hi,

I am loading a 4-band image using io.imread, but I am having problems to visualize the true color (R,G,B) and Composite Infrared (IR,R,G) images.

I have try different band combinations, but I couldnt view the right visualization (see attached image)

Any suggestion?

Thanks in advance, Jaime

[Stéfan van der Walt]

Hi Jaime

Can you tell us what you expect to see? Can you also share an example image?

Thanks
Stéfan

[Jaime Lopez Carvajal]

Hi Stefan,

Of course, I would like to visualize the image like this (attached image), that was obtained using ArcMap software, with the usual band order for True Color (RGB) and Composite Infrared (CIR) in remote sensing applications.

The tiff image that I loaded has 4 bands, so I would think that the order would be: band0(Red), band1(Green), band2(Blue), band3(Infrarred), 

so if I want to visualize the true color image, I would need to use this instruction:

TrueColor_img = img11[:,:,(0,1,2)] 

but as you can see, the obtained visualization is very different from natural color.

I hope this helps, Jaime

[Stéfan van der Walt]

Hi Jaime

On 21 March 2016 at 16:02, Jaime Lopez Carvajal <jalo...@gmail.com> wrote:
> Of course, I would like to visualize the image like this (attached image),
> that was obtained using ArcMap software, with the usual band order for True
> Color (RGB) and Composite Infrared (CIR) in remote sensing applications.
> The tiff image that I loaded has 4 bands, so I would think that the order
> would be: band0(Red), band1(Green), band2(Blue), band3(Infrarred),
> so if I want to visualize the true color image, I would need to use this
> instruction:
> TrueColor_img = img11[:,:,(0,1,2)]

I'll take a look. Can you please upload an example tiff?

Thanks
Stéfan

[Jaime Lopez Carvajal]

Hi Stefan,

Sure, I attached a small fraction of original image, because Remote Sensing images are very big.

The TIFF image has 4 bands.

Hope this helps, Jaime

[Stéfan van der Walt]

Hi Jaime

These are the layers coming out of the TIFF reader:

https://gist.github.com/6881c3527b3099c49891

Can you recognize which ones are which? Do they look correct? Maybe
you can compare them against the ones you see in ArcMap.

Thanks!
Stéfan

> --
> You received this message because you are subscribed to the Google Groups
> "scikit-image" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to scikit-image...@googlegroups.com.
> To post to this group, send email to scikit...@googlegroups.com.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/scikit-image/eb050d76-64f7-4f24-a67f-aa2a4795d1c5%40googlegroups.com.
>
> For more options, visit https://groups.google.com/d/optout.

[Christoph Gohlke]

I think the issue might be that the bands contain uint8 data stored as
uint16 (one extra bit seems used for GDAL's nodata) and the display
function `show_img` does not handle this. In this case casting the RGB
array to uint8 before display might work: show_img(img11a.astype('uint8'))

Christoph

[Jaime Lopez Carvajal]

Hi Stefan,

Accord to ArcMAp, the order of layers is 0,1,2,3 correspond to Red, Green, Blue, Infrared.

This way, if I want to visualize a true color image, I just use layers [0,1,2], in that order, and for CIR image the order would be [4,1,2].

So the order you ask me is: layers [0,1,2,3] = layers [Red, Green, Blue, Infrared]

Thanks, Jaime

[Jaime Lopez Carvajal]

Hi Christoph,

Thanks for your suggestion, I will try that, and will tell you if it works.

Jaime

[Jaime Lopez Carvajal]

Hi Christoph,

You was right, that was the problem, and I can see the visualizations that I want.

See the attached image with results.

Thanks, Jaime

[Jaime Lopez Carvajal]

Hi Stefan,

The problem was about radiometric resolution, because the original image is 16-bit, so I have to convert to 8-bit to visualize correctly.

Keep doing skimage great, I am using it on my research project about object based image classification.

Thanks for your time, Jaime