宽带分裂波束探鱼仪在渔业资源调查中有着广泛应用，其方位角估计精度决定了能否准确确定目标鱼群的方向和位置。为了提高方位估计的精度和效率，提出了一种基于改进的正交数字下变频技术的相位校准算法，对该算法进行了仿真，并基于现场可编程逻辑门阵列（Field Programmable Gate Array，FPGA）实现了该算法；然后采用Vitis HLS平台构建了方位角估计知识产权（IP）核，在提高方位角估计精度的同时降低了浮点计算时延；最后用信号发生器进行了测试。结果显示：模块的方位角估计误差不超过0.05°，整体计算延迟不超过420 ns，逻辑资源消耗不超过45%，与系统指标相符。该研究成果已成功应用于宽带分裂波束探鱼仪中，显著提升了系统精度和性能，并产生了一定的经济效益。

Wideband split-beam fishfinder is widely used in fisheries resource surveys, and the accuracy of its azimuth estimation determines whether the direction and position of target fish schools can be accurately determined. In order to improve the accuracy and efficiency of azimuth estimation, a phase calibration algorithm based on improved orthogonal digital down-conversion technology was proposed, simulated, and implemented using the Field Programmable Gate Array (FPGA). An azimuth estimation intellectual property (IP) core was built on the Vitis HLS platform to improve the accuracy of azimuth estimation while reducing the floating-point computation delay; finally, testing was conducted using a signal generator. The results showed that the azimuth estimation error of the module was no more than 0.05°, the total computation delay did not exceed 420 ns, and the logical resource consumption was less than 45%, consistent with the system specifications. This research achievement has been successfully applied in the broadband split-beam fishfinder, significantly enhancing system accuracy and performance and generating certain economic benefits.