Performance evaluation of FPGA based embedded ARM processor for ultrasonic imaging

Abstract

This study evaluates the performance of an FPGA based embedded ARM processor system to implement signal processing for ultrasonic imaging and nondestructive testing applications. FPGA based embedded processors possess many advantages including a reduced overall development time, increased performance, and the ability to perform hardware-software (HW/SW) co-design. This study examines the execution performance of split spectrum processing, chirplet signal decomposition, Wigner-Ville distributions and short time Fourier transform implementations, on two embedded processing platforms: a Xilinx Virtex-5 FPGA with embedded MicroBlaze processor and a Xilinx Zynq FPGA with embedded ARM processor. Overall, the Xilinx Zynq FPGA significantly outperforms the Virtex-5 based system in software applications.