摘要
在复杂环境下,雷达系统面临多目标密集分布、目标轨迹交叉以及散射强度起伏大等挑战,传统基于门限和滤波的检测方法难以同时保证检测精度与目标关联的鲁棒性,文中提出一种基于端到端实例分割网络的目标检测与关联一体化方法。该方法将连续脉压信号转化为时–距二维图像,利用编码–解码结构的神经网络实现像素级目标检测和嵌入特征提取;在训练过程中引入背景约束与类内–类间嵌入损失,增强特征的可分性;推理阶段结合无监督聚类实现多目标的自动区分与关联。实验结果表明,该方法在目标密集、低信噪比和强散射起伏等场景下明显优于传统方法,在检测准确率、目标区分能力与实时性方面均表现出色,为雷达目标智能检测与航迹起始提供了切实可行的智能化建模路径。
Radar systems operating in complex environments face challenges such as dense target distribution,trajectory overlap,and fluctuating scattering intensity.Traditional threshold and filter-based methods often fail to achieve both high detection accuracy and reliable target association.An end-to-end instance segmentation network is proposed to jointly address target detection and association.Pulse-compressed signals are converted into time-range images,and an encoder-decoder network is employed to perform pixel-level detection and feature embedding.Background constraints and embedding losses are incorporated during training to improve feature separability,while unsupervised clustering in the embedding space enables automatic target distinction and association during inference.Experimental results show that the proposed method achieves superior accuracy,robustness,and efficiency compared with conventional approaches,providing a practical framework for intelligent radar detection and track initiation.
作者
申伦豪
余继周
叶春茂
胡鹏飞
王洪淼
SHEN Lunhao;YU Jizhou;YE Chunmao;HU Pengfei;WANG Hongmiao(Beijing Institute of Radio Measurement,Beijing 100854,China)
出处
《现代雷达》
北大核心
2025年第10期85-91,共7页
Modern Radar
关键词
脉压图像
端到端实例分割
多目标检测
目标关联
嵌入特征
检测与关联一体化
pulse-compressed image
end-to-end instance segmentation
multi-target detection
target association
embedding features
joint detection and association
作者简介
申伦豪,男,1994年生,博士,高级工程师,研究方向为雷达目标检测与识别、雷达图像智能处理;余继周,男,1977年生,博士,研究员,研究方向为雷达目标识别、雷达气象探测与多源信息融合等;叶春茂,男,1981年生,博士,研究员,研究方向为空间目标雷达观测、ISAR成像与姿态估计、目标识别等;胡鹏飞,男,1998年生,硕士,工程师,研究方向为雷达系统设计、气动目标识别;王洪淼,男,1994年生,博士,工程师,研究方向为雷达图像处理与目标识别、极化SAR。
