| Summary: | Synthetic Aperture Radar (SAR) images provide important information about our living
earth. However, there are problems associated with the storage and transmission of that
data that are critical to extending their potential applications. To solve this problem we
must find a suitable compression approach that can significantly decrease the volume of the
data without losing any useful information that SAR images may provide.
There are two main parts to this thesis. The first part is concerned with single channel
SAR image compression. Here, single channel represents single polarization, which is used
to obtain images, since for single channel SAR image compression, only one image is used,
and this image is the amplitude image. We focus our investigation on transform coding,
which is a very popular data compression approach; the particular transform we are
interested in is the Discrete Wavelet Transform (DWT). We want to find a way to improve
the DWT based compression method to make it more suitable for SAR image compression.
Based on experimental results, we find out that our goal can be achieved by adaptively
adopting techniques, such as wavelet packet and block coding.
The second part of the thesis involves the investigation of the compression of
multipolarization SAR images, which include three intensity images and two phasedifference
images as a whole data set. The compression method we are concerned with is
called the Principal Component Analysis (PCA), a standard compression technique for
hyperspectral image compression. Our experiment results show that PCA is less efficient
for multipolarization image compression than for multiple hyperspectral images. This is
because PCA is more efficient at compressing multiple images with higher correlation,
which is not true for multipolarization SAR images. In this thesis, we suggest
multipolarization SAR images should be compressed separately to achieve the best
compression performance, instead of grouping them together. === Applied Science, Faculty of === Electrical and Computer Engineering, Department of === Graduate
|
|---|
