Applied Science
Satellite Remote Sensing
Welcome to Serenetics Remote Sensing solution. This introduction
outlines the experience of our Image Analysis capability
in the Remote Sensing field.
Many of our team members were involved in providing cloud
mask for NASA Earth Observing System instruments. The problem
is to identify the presence or absence of clouds in image
pixels obtained from spaceborne instruments. This seemingly
simple task for our human vision is rather difficult to
train a computer system to perform consistently since the
shapes of clouds are not well defined and their transparency
in an image is never constant.
Using image analysis techniques combining both spectral
and spatial characteristics derived from a multi-channeled
image and artificial intelligence methods such as neural
networks, we are not only able to distinguish between clear
and cloudy pixels but also identify terrestrial and atmospheric
features of interest and classify each pixel of an image
into predefined categories. The performance of this classification
system being over ninety percent correct approaches that
of a human expert.
In the process of solving this problem we have also developed
a software tool, called Interactive Visualizer and Image
Classifier for Satellite (IVICS). IVICS is an advanced visualization
tool with build-in image classifier that provides ability
to view, manipulate and overlay multi-resolution and multi-spectrum
images on different viewing angles in a single session (see
figure 1).
In addition to IVICS highpoints above, IVICS is designed
specifically to address concerns faced in the development
of Remote Sensing applications. These concerns are
Performance
Input data consistency
Easy data integration
Rapid development
Operation efficiency
Architecture
