| Smoothing image by regression in IDL [message #88499] |
Sat, 03 May 2014 02:39  |
Dry in water
Messages: 18
Registered: October 2012
|
Junior Member |
|
|
Hello all,
I'd like to predict NDVI as a function of time using polynomial regression for each land cover separately and integrate them into one image. In other word, I'm going to do temporal smoothing on NDVI image.
I have 2 images; one is time series 8-day NDVI image (26 bands) from Apr to Oct. Another one is Land cover image with 6 classes coded form 1-5. The reason to separate each land cover is just to reduce file size. NDVI image is very big, more than 12GB. Due to big size of image, it is taking very long time for this calculation. Then someone suggested me to use timely average data or spatial subset. Since timely information is my main point, I decided to calculate the regression for each land cover. Anyone to help me?
Thanks in advance
|
|
|
|
|
|
| Re: Smoothing image by regression in IDL [message #88508 is a reply to message #88501] |
Sun, 04 May 2014 14:38  |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
On Sunday, May 4, 2014 12:19:59 PM UTC-4, Dry in water wrote:
> On Saturday, May 3, 2014 5:39:36 PM UTC+8, Dry in water wrote:
>
>> Hello all,
>
>>
>
>> I'd like to predict NDVI as a function of time using polynomial regression for each land cover separately and integrate them into one image. In other word, I'm going to do temporal smoothing on NDVI image.
>
>>
>
>> I have 2 images; one is time series 8-day NDVI image (26 bands) from Apr to Oct. Another one is Land cover image with 6 classes coded form 1-5. The reason to separate each land cover is just to reduce file size. NDVI image is very big, more than 12GB. Due to big size of image, it is taking very long time for this calculation. Then someone suggested me to use timely average data or spatial subset. Since timely information is my main point, I decided to calculate the regression for each land cover. Anyone to help me?
Your problem is so specific that you're unlikely to get a quick and meaningful answer from someone else on the group. Consider: if you don't know what you need, then how can anybody else on the newsgroup know better what you need?
I can only recommend generic advice to make prudent decisions. Consult with colleagues and previously published literature. Try to separate what you want to do - model "something" based on land use type, I guess - from how you will accomplish it - some kind of code written in IDL. This newsgroup is most likely to be able to help on specific questions about IDL.
Best wishes,
CM
|
|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
