I have 423 spot.vgt ndvi from first April 1998 till 31 Dec 2009. I want to know how to stack these NDVI to have one image for time series analysis? Or better to ask how is it possible to analyze a pixle change during these period?
Add a comment
|
2 Answers
You can calculate the VCI index (Vegetation Condition Index) , which is a normalized index that indicates the percentage of change of the difference between the current NDVI index and historical NDVI time series minimum with respect to the NDVI dynamic range.
The formula for VCI is :
where:
- NDVI(x,y) is the current year NDVI value at location (x, y)
- NDVI_min,(x,y) is the historically minimum value at location (x,y)
- NDVI_max,(x,y) is the historically maximum value at location (x,y)
VCI therefore tell you the deviation of each pixel from your historically NDVI values.
You can calculate the difference of each pixel with regard to the previous year value using the change Ratio to Previous NDVI index (RPNDVI):
where:
- NDVI_(x,y) is the current year NDVI value at location (x, y)
- NDVI_i-1,(x,y) is the previous year NDVI value at location (x,y)
A twist of the above index, is change Ratio to Mean NDVI (RMNDVI) which calculates the percentage change of each year not based on the previous year, but the mean value of the pixel calculated over n years:
where:
- NDVI_i,(x,y) is the current year NDVI value at location (x, y)
- NDVI_mean,(x,y) is the mean NDVI value at location (x,y) calculated over n years.
Also, you check VegScape website for a of those indexes.
The Composite Bands (Data Management) tool is the one you are looking for. Regarding the time series analysis, there is a great discussion on just that subject here on GIS.SE and here too.
