[GRASS-user] Land use/cover change detection with GRASS GIS

[Nikos Alexandris]

fatih:

> I plan to determine land use/cover changes using different Landsat images
> in GRASS GIS. Is there anybody who has tried this before? I am not
> familiar with the GRASS GIS. So, a document/manual which includes step by
> step description of image processing may help.

GRASS provides the tools to do it :-). What you really need is to decide which 
methods to try and finally use (worth reading [1][2][3]).

 A few more words

Among the most common change detection techniques are:

1. post-classification comparison (of a pre and a post data set): classify the 
pre and the post data separately (check the wiki-page that Daniel posted in 
the other reply to your post), then subtract to produce a change map.

2. PCA-based method(s) where PCA is utilised in order to "isolate" the changes 
in less dimensions than the original (given we talk about multi-dimensional 
data sets, e.g. multi-spectral bands of a satellite image acquisition). [4][5]

One PCA-based example is to treat all (pre and post) bands as one data set and 
perform PCA. Changes are expected to appear "isolated" in a few (lower order) 
components. You can use the "enhanced" components to classify the changes.

However, it is required to carefully check the numbers in order to best 
identify the changes of your interest. Visually controlling the outcomes is 
not enough! The eigen-vectors which "compose" the outcomes and the eigen-
values which describe how much of the original variance holds each component.

Have (also) a look here: 
<http://geoinformatics.fsv.cvut.cz/gwiki/Change_Detection_with_GRASS_GIS_-
_Comparison_of_images_taken_by_different_sensors>.

Cheers, Nikos

---
[1] Lu, D and Mausel, P and Brondizio, E and Moran, E, "Change detection 
techniques", International Journal of Remote Sensing, vol. 25, no. 12, pp. 
2365, 2003.

--%<--- PCA-based
[2] Deng, JS and Wang, K and Deng, YH and Qi, GJ, "PCA-based land-use change 
detection and analysis using multitemporal and multisensor satellite data", 
International Journal of Remote Sensing, vol. 29, no. 16, pp. 4823, 2008.

[3] Koutsias, N and Mallinis, G and Karteris, M, "A forward/backward principal 
component analysis of Landsat-7 ETM+ data to enhance the spectral signal of 
burnt surfaces", ISPRS Journal of Photogrammetry and Remote Sensing, vol. 64, 
no. 1, pp. 37, 2009.
--->%--

[4] <http://grass.osgeo.org/grass64/manuals/html64_user/i.pca.html>

[5] an attempt to collect details on using PCA (in GRASS, R): 
<http://grass.osgeo.org/wiki/Principal_Components_Analysis>