基于梯度矢量特征的毫米波雷达图像滤波方法

Image Filtering Method for Millimeter-Wave Radar Based on Gradient Vector Features

伴随着我国低空经济发展进入快速扩张期,低空航空器飞行安全问题成为焦点,其中电力线网带来的挑战尤为突出。针对电力线网的图像检测问题,本文基于电力线网的点(塔)-线分布特点和毫米波雷达前视成像结果,提出适用于该场景下的毫米波雷达图像滤波方法——改进联合双边滤波。该方法通过计算雷达图像梯度值得到梯度矢量核,与空间核相乘得到卷积模板,从而实现检测区域增强。经外场挂飞实验验证,该方法可以有效保护目标边缘,显著提高信噪比。

With the rapid expansion of low-altitude economy in China,the issue of flight safety for low-altitude aircraft has become a focus,especially the challenges posed by power line networks.To address the image detection problem of power line networks,this paper proposes a millimeter-wave radar image filtering method based on the point(tower)-line distribution characteristics of power line networks and the forward-looking imaging results of millimeter-wave radar.The improved joint bilateral filtering method is suitable for this scenario.The method calculates the gradient value of radar images to obtain the gradient vector kernel,which is multiplied by the spatial kernel to obtain the convolution template,thus achieving enhanced detection area.The field flight experiment verifies that this method can effectively protect the target edges and significantly improve the SNR.