JDAF & IMAGEbot
TOPS is implemented
using a flexible framework that enables fast and easy integration of
new models and data streams into an automated system. The core
components of this framework are JDAF and the IMAGEbot planner. JDAF
consists of a large set of data processing and image analysis
algorithms that are deployed to pre-process and post-process inputs and
outputs of the TOPS ecosystem models. When we want to process new data
with our existing models, we re-use the JDAF algorithms to create
intermediate datasets that adhere to the model's input specifications
so that we can execute our models without having to alter the science
implementation. Because the pre-processing itself can be a very
complex process, involving for example data acquisition, mosaicking,
reprojection, subsetting, scaling etc., we have developed a
planner-based agent (IMAGEbot), which automatically generates the
sequence of processing steps needed to perform the appropriate data
transformations.
In
other words, JDAF provides all the processing components of the system
and IMAGEbot determines how they fit together to achieve the desired
goal, creates a plan, and executes it. This gives a great flexibility
to the TOPS software and speeds-up significantly the integration of
both datasets and models into new applications. Additionally, JDAF
provides interface to the database system and to web services
capabilities for seamless access to both data and services provided by
TOPS.
As
currently deployed within TOPS, JDAF and ImageBot perform two dynamic
functions critical for the real-time monitoring, modeling and
forecasting of ecosystem conditions: gridding of weather observations
to create continuous fields of climatic parameters, and acquisition and
processing of satellite data for initializing or verifying the models.
;){kind=link}