
Fig. 1 An analysis of trends in sea surface temperature from 1982 to 2006.
|
The Earth Trends Modeler
The Earth Trends Modeler is specially designed for the analysis of
image time series from earth observing systems. It includes a
coordinated suite of data mining tools for the extraction of trends
and underlying determinants of variability, and will be of special
importance to scientists focused on climate change and ecosystem dynamics.
|
|

Fig. 2 SEGMENTATION module creates an image of segments that have a spectral similarity across many input bands.

Fig. 3 SEGTRAIN assigns these segments to specific land cover types for the development of training site data.
Fig. 4 SEGCLASS classifies the imagery using a majority rule algorithm to assign each segment to a class. SEGCLASS can improve the accuracy of a pixel-based classification and produce a smoother map-like classification result while preserving the boundaries between the segments.
|
Segmentation
IDRISI Taiga provides three new modules for classification from image
segments. Segmentation is a process by which pixels are grouped that
share a homogeneous spectral similarity. Across space and overall input
bands, a moving window assesses this similarity and segments are defined
according to a stated similarity threshold. The three modules provide
a sequence for segment-based classification. SEGMENTATION creates an
image of segments. SEGTRAIN interactively develops training sites
and signatures. SEGCLASS classifies the imagery utilizing a majority
rule algorithm.
Land Change Modeler Enhancements
IDRISI's Land Change Modeler now includes an interface to MARXAN that
facilitates its use. Given a number of potential reserve sites, MARXAN
identified a portfolio of sites which meets a particular target.
The Land Change Modeler application also now includes a validation panel.
A 3-way cross-tabulation can be run between the later land-cover map,
the prediction map, and a map of reality in order to evaluate the
result of a possible prediction outcome. Hits, misses and false alarms
are reported.
Additionally, options to print and manipulate the graphs generated by
Land Cover Modeler are now included.
Display Navigation Tools
Enhancements have been made to the pan and zoom in/zoom out functions.
The stretch options on Composer have also been revised.
Other Features
IDRISI Taiga also includes an extension of the Multi-Layer Perception
neural network classifier to support multiple regression applications,
an ISODATA unsupervised classification procedure, and additional time
series utilities.
A wide range of new import procedures have been added. Many of these
are designed to support the varied earth observing system image time
series. In addition, the Taiga Edition of IDRISI now supports the
creation of KML layers, both as single layers and as pyramid structures
for streaming over the Internet. Finally, an interface is provided to
the open source GDAL raster translation software.
|