General Description

The SPIRITS software is a stand-alone image analysis toolbox developed by VITO for the Joint Research Centre of the European Commission. Although the software works on any type of raster data, it focuses on time series analysis of daily, ten-daily or monthly composites from low resolution sensors such as SPOT-VEGETATION, NOAA/METOP-AVHRR, AQUA/TERRA-MODIS, MSG-SEVIRI as well as with gridded meteorological data.
It offers an integrated and flexible analysis environment with a user-friendly graphical interface and the possibility to automate complex processing chains. Particularly, it provides a large number of time series analysis tools to:

  • perform spatial and temporal processing steps on image time series
  • compute vegetation anomalies as compared to reference years or long term averages
  • extract aggregated statistics for vegetation related indicators
  • generate graphical outputs such as maps and graphs for analysts

The software is extensively documented. It has a User Manual that contains general information on image formats and general concepts on SPIRITS processing and where all SPIRITS functionalities are described in detail. The Tutorial is intended as individual training material. A default sample training data set is provided with the tutorial.

Examples of application

Long term averages and vegetation status anomalies are computed after import of the time series, extraction of the study area and smoothing. Once the map layout is defined, the temporal set of maps is automatically produced. NDVI anomaly images show the effect of the 2010/2011 drought in the Horn of Africa.Heatmap of Africa
Temporal profiles of NDVI and rainfall are generated over administrative regions and crop areas. Once a graph layout showing the temporal profiles of the two variables against their long term statistics is defined, the graphs for all administrative unit and land cover class combinations are automatically generated.
The actual and average start and end of season is determined after import, extraction of the study area and smoothing. During the crop season cumulated FAPAR is calculated and anomalies compared to the average situation are identified.
Image import convert any image format supported by the GDAL library to the SPIRITS standard (ENVI with enriched header file).

Processing modules overview

Import and export

Image import convert any image format supported by the GDAL library to the SPIRITS standard (ENVI with enriched header file)
Vector to raster rasterize ESRI Shapefiles to a user-defined grid
File rename rename file time series and include the image dates in the file names
Image export convert SPIRITS images to IDRISI and ESRI formats or to a 3D ENVI layer stack

Spatial operations

ROI extraction extract a region of interest (ROI) from input images
Resampling resample input images to a user-defined resolution and framing
Generation of AFIs starting from a high-resolution classification, derive low resolution Area Fraction Images (AFIs) with the area fraction of each class per pixel
Low-pass filters spatially smooth input images using moving window filtering

Thematic operations

Rescale rescale, reclass or modify the original image values, change data type
Index compute vegetation indices, difference images and anomalies
Masking mask image pixels by defined intervals or by values in another image
Flagging apply the information of a status mask image (e.g. water, cloud, snow) as flags onto another image
DMP/NPP derive DMP or NPP from FAPAR images and meteorological information
Clustering apply unsupervised classification using an enhanced iso-clustering algorithm

Time series operations within a single year

Smoothing detect noisy observations and smooth signals based on the SWETS algorithm
Compositing compute multi-temporal composites with different frequencies
Averaging derive mean values of image values in a time series between two given dates
Phenology detect the season start and end dates
Pheno averaging compute mean or sum over a time series between the start and end of the season

Time series analysis
over the years

Long-term statistics compute long-term stats (mean/min/max/sd/…) on a multi-annual time series
Anomalies compare actual images with the corresponding long-term statistics
Similarity analysis compare a single year with others to detect the most similar year, or compare a single year with the long-term average to define the overall advance or delay
Similarity-based yield assessment assess crop yields based on the assumption that the yield of the current year will mostly resemble the yield of the most similar year

Analysis tools

Map composer compose simple GIS maps and store as templates to apply on a time series
Extraction of statistics extract statistics aggregated by administrative and thematic areas and upload to a database
Graph composer browse and query the database to show statistics in graphs and store favorite graphs as templates to apply automatically to other regions or land use types
User tools run external programs in a SPIRITS GUI and create processing chains by concatenating different programs