A virtual multi-channel based data compression method for high resolution synthetic aperture radar data
Chenglin Sun, Qi Ming, Shichao Bu, Junli Chen · IET conference proceedings. · 2026
Aiming at the under-utilisation of frequency-domain redundancy and the rigidity of bit-allocation strategy in the traditional block adaptive quantisation (BAQ) algorithm in the compression of satellite-borne synthetic aperture radar (SAR) raw data, this paper puts forward a high-resolution SAR data compression method based on virtual multi-channel. The compression efficiency and fidelity are significantly improved by azimuthal virtual multichannel construction and joint dynamic bit allocation in the frequency domain. The core innovations include: (1) reconfiguring the single-channel azimuthal data into virtual multichannel form to simulate the frequency-domain energy distribution characteristics of multichannel SAR systems; (2) performing the azimuthal discrete Fourier transform (DFT) on the virtual multichannel data to generate the multibeam frequency-domain data, and optimizing the quantization strategy by using the frequency-band energy concentration characteristics; (3) proposing a joint dynamic bit allocation algorithm based on the variance and the band energy to achieve the globally optimal compression efficiency. algorithm to achieve globally optimal compression. The experimental results show that, compared with the traditional BAQ algorithm, the signal quantisation noise ratio (SQNR) is improved by about 1.77 dB under the same compression ratio, which verifies the effectiveness of the method.