胡译夫,周剑雄,胡卫东,等. 距离像先验引导的扩展目标检测方法[J]. 信号处理,2024,40(3): 587-598. DOI: 10.16798/j.issn.1003-0530.2024.03.016.
引用本文: 胡译夫,周剑雄,胡卫东,等. 距离像先验引导的扩展目标检测方法[J]. 信号处理,2024,40(3): 587-598. DOI: 10.16798/j.issn.1003-0530.2024.03.016.
citation‍:‍HU Yifu,ZHOU Jianxiong,HU Weidong,et al. An extended target detection method guided by range profile prior[J]. Journal of Signal Processing,2024,40(3):587-598. DOI: 10.16798/j.issn.1003-0530.2024.03.016.
Citation: citation‍:‍HU Yifu,ZHOU Jianxiong,HU Weidong,et al. An extended target detection method guided by range profile prior[J]. Journal of Signal Processing,2024,40(3):587-598. DOI: 10.16798/j.issn.1003-0530.2024.03.016.

距离像先验引导的扩展目标检测方法

An Extended Target Detection Method Guided by Range Profile Prior

  • 摘要: 扩展目标检测通常采用距离像能量积累检测的方法,由于距离像信息掌握不完备,陷落损失会降低检测性能。本文提出一种距离像先验引导的扩展目标检测方法,通过利用距离像包络模先验,对信号进行积累以提升检测性能。该方法考虑了复距离像与复高斯白噪声的相干叠加与相位预测不准的因素,采用将观测数据取模的检测模型,基于似然比检测(Likelihood Ratio Test,LRT)理论推导了低信噪比下的特征平方匹配检测器。该检测器将目标复距离像的包络模与观测数据的包络模进行平方匹配,并通过门限判决来判断目标是否存在。包络模先验的获取是通过从ISAR图像提取二维散射中心,向对应姿态角下的雷达视线方向进行投影,来获得目标近似的一维散射中心模型,再由该模型进一步生成目标距离像的包络模先验。同时,本文从理论与实验两方面分析了能量检测器和特征平方匹配检测器之间的关系,通过散射中心模型重构与暗室测量的方法获取数据进行了实验验证。实验结果表明:在低信噪比下,距离像先验引导的特征平方匹配检测器能有效提升目标的检测性能,并且对先验模型失配的情况具有良好的适应性。

     

    Abstract: ‍ ‍Extended target detection often employs a range-profile energy-accumulation detection method. However, incomplete range-profile information often reduces the detection performance as a result of a collapsing loss. To address this issue, this paper proposes an extended target detection method that is aided by a prior range profile, which effectively utilizes the prior envelope model of the range profile to accumulate the signal and improve the detection performance. The proposed method takes into account the coherent superposition of the complex range profile and complex Gaussian white noise, as well as the inaccurate phase prediction. It uses a detection model that takes the modulus of the observed data, and based on the likelihood ratio detection theory, derives a feature square matching detector under a low signal-to-noise ratio. This detector matches the envelope model of the target complex range profile with the envelope model of the observed data using square matching, and uses a threshold judgment to determine the presence of the target. The acquisition of the prior range profile is achieved by extracting the two-dimensional scattering center from the ISAR image and projecting it in the radar line-of-sight direction under the corresponding posture angle to obtain an approximate one-dimensional scattering center model of the target. The envelope model of the target distance image is then generated based on this model. Furthermore, this study analyzed the relationship between the energy detector and feature square matching detector from both theoretical and experimental perspectives. An experimental verification was conducted using data obtained through the reconstruction of the scattering center model and darkroom measurement methods. The experimental results demonstrated that the prior range profile-aided feature square matching detector could effectively improve the detection performance for targets under a low signal-to-noise ratio and had good adaptability to a prior model mismatch. This approach provides a new way to improve extended target detection performance under certain prior range profile conditions.

     

/

返回文章
返回