Low-bit quantization for UAV SAR data and its complexity analysis

Abstract

Due to the limitation of size and load, an unmanned aerial vehicle (UAV) can only provide limited hardware computing resources. How to obtain high-quality synthetic aperture radar (SAR) imaging with the lowest computational complexity is an important problem in UAV-based SAR system design. In this paper, the disadvantages of existing algorithms are analyzed. Different quantization approaches are used to quantify the raw data of SAR in low bits. An image quality improvement method based on a single frequency threshold is proposed for low-bit quantized SAR data. On this basis, the computational complexity of SAR imaging matched filtering processing under different quantization strategies is studied, and the corresponding imaging quality is quantitively evaluated. This paper provides a reference for achieving the tradeoff between computational complexity and imaging quality in SAR imaging application based on UAV platform.