摘要： 前视声纳通道数目多、采样频率高，获取方位和距离信息的运算量大，对信号实时处理能力要求较高。为此，利用现场可编程门阵列(FPGA)内部运算单元以及并行分布式运算结构，对原始图像中的正交下变频和波束形成算法的FPGA实现进行研究。建立前视声纳信号模型，对核心算法进行理论推导，并介绍FPGA实现方案中的难点，给出资源占用情况。通过水池实验和湖上静态实验，验证该算法具有实时运算的正确性。

关键词: 前视声纳, 多通道高采频, 正交解调, 波束形成, 现场可编程门阵列, 并行运算

Abstract: The characteristics of multiple channel and high sample frequency in forward looking sonar demands a heavy computation burden for azimuth and distance information processing, which sets a higher requirement for the real time computation. By the processing units and distributed operation in Field Programmable Gate Array(FPGA), the bottle-neck of speed can be resolved on a large scale. The paper mainly studies the implementation of quadrature demodulation and beam-forming algorithm applied for the original imaging information, based on FPGA. The signal model is established, and the core algorithm is derived. The difficultness in the implementation is presented, and the utilization of FPGA resources is provided. A validation of the processing schemes is conducted with real data in the tank and lake yielding accurate results.

Key words: forward looking sonar, multiple channel high sample frequency, quadrature demodulation, beamforming, Field Programmable Gate Array(FPGA), parallel computation

中图分类号: 

• TP39