This paper presents a new method of High Resolution Range (HRR) profile formation based on Linear Frequency Modulation (LFM) signal fusion of multiple radars with multiple frequency bands. The principle of the multipl...This paper presents a new method of High Resolution Range (HRR) profile formation based on Linear Frequency Modulation (LFM) signal fusion of multiple radars with multiple frequency bands. The principle of the multiple radars signal fusion improving the range resolution is analyzed. With the analysis of return signals received by two radars,it is derived that the phase difference between the echoes varies almost linearly with respect to the frequency if the distance between two radars is neg-ligible compared with the radar observation distance. To compensate the phase difference,an en-tropy-minimization principle based compensation algorithm is proposed. During the fusion process,the B-splines interpolation method is applied to resample the signals for Fourier transform imaging. The theoretical analysis and simulations results show the proposed method can effectively increase signal bandwidth and provide a high resolution range profile.展开更多
基于视频的目标检测在恶劣天气情况下识别效果较差,故需弥补视频缺陷、提高检测框架的鲁棒性。针对此问题,文中设计了一个基于雷达和视频融合的目标检测框架,利用YOLOv5(You Only Look Once version 5)网络获得图片特征图与图片检测框,...基于视频的目标检测在恶劣天气情况下识别效果较差,故需弥补视频缺陷、提高检测框架的鲁棒性。针对此问题,文中设计了一个基于雷达和视频融合的目标检测框架,利用YOLOv5(You Only Look Once version 5)网络获得图片特征图与图片检测框,利用基于密度的聚类获得雷达检测框,并将雷达数据进行编码,得到基于雷达信息的目标检测结果。最后将两者的检测框叠加得到新ROI(Region of Interest),并得到融合雷达信息后的分类向量,提高了在极端天气下检测的准确率。实验结果表明,该框架的mAP(mean Average Precision)达到了60.07%,且参数量仅为7.64×10^(6),表明该框架具有轻量级、计算速度快、鲁棒性高等特点,可以被广泛应用于嵌入式与移动端平台。展开更多
雷达信号融合成像是一种能显著提高成像分辨率的参数化新方法.基于改进的Root-Music的传统融合方法对噪声敏感,且存在模型极点失配的问题.本文通过将MEMP(Matrix Enhancement and Matrix Pencil)的二维频率估计方法推广到稀疏数据域,提...雷达信号融合成像是一种能显著提高成像分辨率的参数化新方法.基于改进的Root-Music的传统融合方法对噪声敏感,且存在模型极点失配的问题.本文通过将MEMP(Matrix Enhancement and Matrix Pencil)的二维频率估计方法推广到稀疏数据域,提出了一种基于扩展矩阵增强矩阵束(EMEMP)的融合新方法.此方法首先构造每一维联合增强矩阵,使其满足MEMP算法的配对要求,然后利用MEMP方法估计模型极点,进行极点配对,然后估计模型系数,最后内插频谱以达到融合的目的.实验结果表明相对于传统融合方法,该方法解决了极点失配的问题,提高了模型参数估计的稳健性.展开更多
针对雷达信号脉内调制识别算法存在着准确率低的问题,提出一种新的雷达脉内调制类型自动识别方法,该方法首先提取雷达信号时频图像的形状特征和纹理特征构成融合特征,然后将融合特征输入随机森林分类器,实现信号的分类识别。仿真实验中...针对雷达信号脉内调制识别算法存在着准确率低的问题,提出一种新的雷达脉内调制类型自动识别方法,该方法首先提取雷达信号时频图像的形状特征和纹理特征构成融合特征,然后将融合特征输入随机森林分类器,实现信号的分类识别。仿真实验中对8种常见的不同调制类型的雷达信号进行识别,提出的算法在信噪比为-2 d B时识别准确率可以达到90%以上,验证了该方法的有效性。展开更多
文摘This paper presents a new method of High Resolution Range (HRR) profile formation based on Linear Frequency Modulation (LFM) signal fusion of multiple radars with multiple frequency bands. The principle of the multiple radars signal fusion improving the range resolution is analyzed. With the analysis of return signals received by two radars,it is derived that the phase difference between the echoes varies almost linearly with respect to the frequency if the distance between two radars is neg-ligible compared with the radar observation distance. To compensate the phase difference,an en-tropy-minimization principle based compensation algorithm is proposed. During the fusion process,the B-splines interpolation method is applied to resample the signals for Fourier transform imaging. The theoretical analysis and simulations results show the proposed method can effectively increase signal bandwidth and provide a high resolution range profile.
文摘基于视频的目标检测在恶劣天气情况下识别效果较差,故需弥补视频缺陷、提高检测框架的鲁棒性。针对此问题,文中设计了一个基于雷达和视频融合的目标检测框架,利用YOLOv5(You Only Look Once version 5)网络获得图片特征图与图片检测框,利用基于密度的聚类获得雷达检测框,并将雷达数据进行编码,得到基于雷达信息的目标检测结果。最后将两者的检测框叠加得到新ROI(Region of Interest),并得到融合雷达信息后的分类向量,提高了在极端天气下检测的准确率。实验结果表明,该框架的mAP(mean Average Precision)达到了60.07%,且参数量仅为7.64×10^(6),表明该框架具有轻量级、计算速度快、鲁棒性高等特点,可以被广泛应用于嵌入式与移动端平台。
文摘雷达信号融合成像是一种能显著提高成像分辨率的参数化新方法.基于改进的Root-Music的传统融合方法对噪声敏感,且存在模型极点失配的问题.本文通过将MEMP(Matrix Enhancement and Matrix Pencil)的二维频率估计方法推广到稀疏数据域,提出了一种基于扩展矩阵增强矩阵束(EMEMP)的融合新方法.此方法首先构造每一维联合增强矩阵,使其满足MEMP算法的配对要求,然后利用MEMP方法估计模型极点,进行极点配对,然后估计模型系数,最后内插频谱以达到融合的目的.实验结果表明相对于传统融合方法,该方法解决了极点失配的问题,提高了模型参数估计的稳健性.
文摘针对雷达信号脉内调制识别算法存在着准确率低的问题,提出一种新的雷达脉内调制类型自动识别方法,该方法首先提取雷达信号时频图像的形状特征和纹理特征构成融合特征,然后将融合特征输入随机森林分类器,实现信号的分类识别。仿真实验中对8种常见的不同调制类型的雷达信号进行识别,提出的算法在信噪比为-2 d B时识别准确率可以达到90%以上,验证了该方法的有效性。