I linked to a repo in GitHub that may or may not be helpful. Not entirely sure. Good luck though, I need to figure out how to do this as well. 

https://github.com/azvoleff/Chitwan_MODIS_SVIs

Dear Sofia:

Follow this Links with valuable informatin about interpolation, may it be useful for your work:

http://www.r-bloggers.com/spline-interpolation-of-temporal-resolution-for-satellite-images/

https://github.com/mkao006/interpolate_ndvi_satellite_image

Regards!

Many thanks!

You can easily run the Timesat software by installing the matlab compiler in windows os. IF you follow the instructions is not that difficult - it has a gui that you can easily handle if you follow the tutorial. If you process your data correctly then Timesat can extract the metrics you want

I recommend the use of Phenosat which can be obtained from the address.

http://www.fc.up.pt/PhenoSat/software.html.

I also recommend the paper: Monitoring Vegetation Dynamics Inferred by Satellite Data Using the PhenoSat Tool

https://www.researchgate.net/publication/234079317_Monitoring_Vegetation_Dynamics_Inferred_by_Satellite_Data_Using_the_PhenoSat_Tool

Article Monitoring Vegetation Dynamics Inferred by Satellite Data Us...

As Jan pointed, MODIS package is very useful in data preprorcessing. May be phenex package (auxiliary functions for phenological data analysis) is useful in extraction of the parameters you are looking for:

http://cran.r-project.org/web/packages/phenex/index.html

Cheers, Mohammad

greenbrown package in R could be of great help

How about phenofit (https://github.com/kongdd/phenofit)?

Hi Sofia,

if you need the NDVI values from specific regions of your map you would need to create a .shp file containing some polygons that have been placed over your region of interest. This can be done by using a standard GIS tool (e.g. QGIS).

Once you have your shapefile (.shp) and your image (raster) both can be loaded into R by using the raster package. After loading your raster image (which consists of multiple bands) you can use it just a you would use a dataframe and calculate any index you like (e.g. NDVI):

ndvi

PhenologyRaster {greenbrown} may help you.

Does any of these methods capable to derive multiple seasons within one cycle?

Hello,

I want to extract seasonal metrics of precipitation, evapotranspiration and vegetation indices in TIMESAT, for a long series of 18 years. However, I'm doing a one-year test to see if it runs. I converted the 23 images into the ENVI (BSQ) format, and created the textlist file to input the dataset into TIMESAT.

In the window that generates the curves, it is giving the following error:

Warning - too many missing data points or constant data Pixel will be excluded - check data range and amplitude.

How can you help me, please?

Anacleto Diogo You need to mask pixels with no seasonality first.

Dongdong Kong

Hello,

Thanks for answering. I intend to use for a series of 18 years (2001-2018). I tried to test in TIMESAT, only three years, to see if it works. I'll explain the output I used: I downloaded the dataset of the MOD13 (Geotiff) time series, already cut to my study area (500,000 km²), approximately. I plan to work with EVI. I saved each EVI image (69 images-three years for testing) in the ENVI BSQ format. I created the txt files (readme and list) for redirection. And showed the error I presented to you.

Given this procedure I used, what am I doing wrong? What is missing to implement?

Can you explain to me the correct steps that I must follow.

Thanks.

Jiu Feng recommendation is probably the best choice because it handles the per-pixel analysis automatically. If you want to try more fitting functions and threshold methods you can have a look at the phenopix R package (https://cran.r-project.org/web/packages/phenopix/phenopix.pdf).

Hugo Bendini as far as I know there is no off the shelf function to handle multiple peaks in a single season.