RE: [erdas-l] stretching imagery using imagine and saving LUT for batch processing

[Adrian Benitez]

Hi Judy,

For digital analysis, usually, the digital procedure manages automatically the range and distribution of the datas.
So, specialists recommend do not transform the original digital values.

For visual analysis, the right type of stretch depends on the Objective and the Imagery under analisys. Linear stretching are non aggressive and it keeps the linearity of the datas. On the other hand, non linear stretching (equalization, gamma,etc) produces images who shows the maximum contrast, but loosing the relative differences between the features.

There are no bads or goods, just appropriate methods. For example, you can use a linear stretch in a first step (taking in account the individual statistic of each image, avoiding to include lakes, the sea, and clouds in the stats.) in order to get some color homogenization. Then, you can perform non linear stretch in an specifics areas where you want to enhance the image to better understand vegetation differences.

Regards,

Adrian E. Bรญtez, Ing. Agr.
Responsable Geographic Imaging, Aeroterra S.A.
aben...@aeroterra.com www.aeroterra.com
Argentina  


-----Mensaje original-----
De: erd...@lyris.dundee.net [mailto:erd...@lyris.dundee.net] En nombre de Judith Brandt
Enviado el: viernes, 13 de febrero de 2009 10:32 p.m.
Para: ERDAS Discussion List
Asunto: [erdas-l] stretching imagery using imagine and saving LUT for batch processing

Hello all,

We are wanting to use imagine to apply a stretch to some 4-band imagery collected by a digital mapping camera. We are starting with imagery that has not been color or tone balanced and no stretch has been applied to it. Histogram range is a bit... narrow. We would like to enhance the radiometric info by applying a "gamma stretch" -- we think. We basically want to use the magery for some vegetation analysis. Is a "gamma stretch" a standard equation that we can apply to the values? I have searched the imagine help and have found nothing on it there. Please let me know if you know of a gamma stretch or if this particular steretch is even the right type of steretch for this case. thanks!


Judy Brandt
GIS Analyst
Southern Nevada Water Authority
Groundwater Resources Department
Data Resources Division
(702) 862-3727

---
You are currently subscribed to erdas-l as: aben...@aeroterra.com.
To unsubscribe send a blank email to leave-31255499-87054609.6468...@lyris.dundee.net

---
You are currently subscribed to erdas-l as: erdas-l-gar...@googlegroups.com.
To unsubscribe send a blank email to leave-31286445-120484586.1d5...@lyris.dundee.net

[Judith Brandt]

At least in one case what we are trying to do is reverse the color balance that was originally applied to the image data. The 4-band CIR image was color balanced by the vendor using a single LUT on many images (on the R, G and B bands). The NIR band was left to do...whatever. This resulted in some very odd looking false color images. Perhaps there is more wrong with the original image data than simply the color balance, but we won't know until we try this color balance reversal. I have LUT values the vendor used to do the original color balance, so I am using those to create an LUT that will perform the reverse. I plug in those lut values as a custom LUT for the "LUT stretch" model. I then use that same LUT model to reverse the rest of the 1,000 or so images and hope for the best.

After that, I am going to have to create my own lut using the Imagine tools to stretch the data, trying to find a stretch that looks closer to "reality" than what I have right now. Then I must take that and apply that same stretch to that large number of images and hope for the best.

Is it possible to use the LUT stretch model that permanently applies a LUT (using a reference image) to the new image? -- I tried doing this before and the internal lut for the new image was created, but the data from the new image was the same as the original. In other words, I had to stretch the data to see the stretch I was trying to apply pemanently. (I would like it to work as it does when custom values are put into the LUT model.)

---
You are currently subscribed to erdas-l as: erdas-l-gar...@googlegroups.com.

To unsubscribe send a blank email to leave-31373246-120484586.1d5...@lyris.dundee.net

[Donn Rodekohr]

Judith,

Making a permanent LUT stretch involves translating the raw input image
pixels to the stretched values.

First manipulate the LUT using the contrast tools of your choice. Apply
the stretch and then SAVE the changes to create a LUT in the .img. This
is a semi-permanent LUT in that the Imagine viewer will always use this
stretch when opening this file .... until you change it using a contrast
tool and save the changes. The LUT is save for each band. the
advantage is that the raw values are unchanged so you can always go back
to the original. If you want to semi-permanently apply this LUT to
another image you can open the contrast tool and export the LUT to a
file name. to apply this LUT to another image open the image, Open the
Contrast tool, import the LUT you just saved, apply it to the 2nd image,
and save it. Your raw values are still in tact but the image will
always display with this stretch, even when sent to the printer.

If you want to alter the raw values to be equal to the stretched values
you can do this with Interpreter - Radiometric Enhancement - Lookup
Table stretch. This program will stretch the input data values through
the LUT to produce an output file with the stretched values. This is
permanent and cannot be undone without going through the painful process
that you described earlier.

If you contractor has the LUT that you are performing the reverse
stretch process, shouldn't they also have the original unstretched
photos? Perhaps it would be easier to get them than try to unstretch a
photo....

Donn Rodekohr, Spatial Analyst
230 Funchess Hall
Agronomy and Soils Department
Auburn University
Auburn, AL 36849
334-844-3989
>>> "Judith Brandt" <judy....@snwa.com> 02/25/09 12:22 PM >>>

You are currently subscribed to erdas-l as: rod...@auburn.edu.

To unsubscribe send a blank email to

leave-31373246-58163005.68a6...@lyris.dundee.net

---
You are currently subscribed to erdas-l as: erdas-l-gar...@googlegroups.com.

To unsubscribe send a blank email to leave-31376292-120484586.1d5...@lyris.dundee.net