Browse Prior Art Database

CODING OF SAR PHASE HISTORY DATA

IP.com Disclosure Number: IPCOM000009713D
Original Publication Date: 2000-Jan-01
Included in the Prior Art Database: 2002-Sep-12
Document File: 3 page(s) / 142K

Publishing Venue

Motorola

Related People

Glen P. Abousleman: AUTHOR

Abstract

To compress SAR phase history data, we pro- pose two methods that arc optimized for different performance criteria. The first is a variable-rate sys- tem that uses entropy-constrained trellis-coded quantization (ECTCQ) in conjunction with adaptive arithmetic encoding. The second system is a fixed- rate design that uses channel-optimized trellis-coded quantization (COTCQ) designed for the binary sym- metric channel.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 50% of the total text.

Page 1 of 3

0 M MOTOROLA Technical Developments

CODING OF SAR PHASE HISTORY DATA

by Glen P. Abousleman

  To compress SAR phase history data, we pro- pose two methods that arc optimized for different performance criteria. The first is a variable-rate sys- tem that uses entropy-constrained trellis-coded quantization (ECTCQ) in conjunction with adaptive arithmetic encoding. The second system is a fixed- rate design that uses channel-optimized trellis-coded quantization (COTCQ) designed for the binary sym- metric channel.

  Synthetic aperture radar (SAR) sensors have extremely high unprocessed data rates. This unprocessed data (also called raw, phase history, or radar echo data) is typically downloaded to a ground station for storage and/or processing. Due to the vast amount of raw data produced, bandwidth con- straints make transmission of the entire data record difficult. These difficulties can be successfully overcome by compressing the raw data such that the fidelity of the generated SAR image is not compro- mised unduly.

  Various methods can be used to compress the raw phase history data. For example, block adaptive quantization (BAQ) [l] divides the phase history data into non-overlapping blocks and normalizes each block to zero-mean and unit variance. A scalar quantizer designed for the zero-mean, unit-variance Gaussian source is then used to quantize each block. Vector quantization (VQ) can also be applied to each block [2]. The vector dimension can be varied, and the generalized Lloyd algorithm is used to "train" the vector codebooks. Typically, separate codebooks are used for the real and imaginary com- ponents of the complex phase history data. Other methods such as tree-searched multi-stage VQ and trellis-coded quantization can also be used.

  There are two types of quantization schemes that can be used for compression. The first scheme produces a variable-length bit stream that is tailored

to the statistics of the encoded source. That is, after quantization, an entropy coding algorithm such as Huffman or Arithmetic encoding is used to assign short codewords to quantization indices whose prob- ability of occurrence is very high, and longer code- words to those quantization indices that have a low probability of occurrence. Although affording high- er signal-to-noise ratio (SNR) performance, entropy- coding schemes suffer from buffer synchronization problems and high susceptibility to channel errors. The second quantization scheme is fixed-rate in nature. Here, all quantization indices are assigned binary codes of equal length (such as the natural binary code). The SNR performance of a fixed-rate system will always be less than that of a properly designed variable-rate system.

  However, a fixed-rate system is much more robust to bit errors introduced by the communica- tion channel.

  To compress SAR phase history data, we pro- pose two methods that are optimized for different performance criteria. The fast is a variable-rate s...