雷达目标三维特征的提取与识别研究

以逆合成孔径雷达(ISAR)成像技术和计算机视觉理论为基础，提出了一套新的从动态目标ISAR成像序列中提取目标散射点三维结构信息，以此作为目标特征的识别方法。这一研究方法包含了4个重要环节：动态目标的ISAR像序列的获得；“散射点像元”质心的检测、跟踪和匹配；基于光流分析的目标散射点三维结构特征的提取；目标三维特征的识别。由于经过了时间和空间上信息的积累，目标的散射点的三维结构特征具有稳定和直观的特点。作为一种新的目标识别的依据是很有效的，且只需要较少的训练样本就可以获得较高的识别率。     

Point cloud-based scene flow estimation on realistically deformable objects: A benchmark of deep learning-based methods

Flow estimation on 3D point clouds is a challenging problem in the field of computer vision, which has great significance in many areas, such as autonomous driving and human interaction applications. Within the last years, the field of motion analysis has made great progress. The evaluation of the existing approaches mostly focuses on scenarios where objects are affected by rigid transformations. However, in many application areas such as gesture recognition or pose tracking, the detection of shape changes is essential and breaking them down to local rigid transformations is accompanied by loss of information. One component of our contributions is that we specifically prepared existing datasets for scene flow estimation on deformable objects. Additionally, we benchmark existing methods and analyze their behavior on various subtasks. The results show that already close to 80% of correct correspondences can be found on synthetic hand data, while only around 50% are found on real hand data. Our experimental validation and analysis help to build an understanding of new possibilities in broader areas. Furthermore, they should help to inspire possible further research directions.     

图象目标识别图形的自适应步级检测

建立图象目标识别模型，用形状、灰度和运动特征描述图象目标。基于目标建模，把目标识别、门限及目标图形区域步级检测、虚漏警调节和目标空域条件有机地结合起来，给出牵引式跟踪系统中图象目标识别图形的自适应步级检测算法。该算法用于检测具有识别特征的图象目标图形，并成功地应用到实际的实时跟踪系统中的图象目标识别和跟踪。对实际图象的处理结果和在实时跟踪系统中的实验说明本文研究的技术的有效性和适用性。     

三维空间中高超声速目标修正三级Hough变换-检测前跟踪算法

针对大型相控阵雷达对于临近空间高超声速目标的检测跟踪问题,该文提出一种修正的3级Hough变换(MTS-HT)检测前跟踪算法。首先,为了解决直接进行3维Hough变换计算量大的问题,将3维点迹依次降维映射至径向距离-时间、方位角-时间和仰角-时间3个平面进行2维Hough变换;其次,为了能够在充分利用点迹能量信息的基础上尽量减小强干扰的影响,在每级Hough变换中采用非相参积累和二值积累相结合的双重积累方式进行点迹筛选;最后,通过引入运动约束以及合并相似航迹的方式进行航迹精简,从而有效降低虚假航迹数,得到按照时序关联的最终检测航迹。仿真结果表明,该算法具有较高的航迹检测概率和较低的虚假航迹检测概率,可以实现对临近空间高超声速目标的有效检测跟踪。     

Weighted Discriminator Function Based Unambiguous Tracking Method for Dual-Frequency Constant Envelope Modulated Signals

The dual-frequency constant envelope modulated (DCEM) signals, represented by alternative binary offset carrier signal, have obvious advantages in code tracking accuracy and anti-multipath performance compared with traditional binary phase shift keying signal, and are attracting more and more attention in the field of global navigation satellite system signal design. However, DCEM signal has a multiple-peaked auto-correlation function, which would lead to the tracking ambiguity problem. To deal with this problem, this paper proposes an unambiguous tracking method based on weighted discriminator function (WDF). This method uses the discriminator functions of noncoherent early minus late power (NELP) and sub carrier phase cancellation (SCPC) to generate a linear WDF without no false-lock points. In addition, a simplified version of the proposed WDF method is also provided for interoperable application. The theoretical code tracking error in thermal noise is derived out and verified. The performance of the proposed WDF is evaluated and compared with SCPC and pseudo correlation function methods with the help of simulation. Results show that our method has a significant improvement in code tracking accuracy. Not only the ambiguity threat is efficiently removed, but also the outstanding performance of DCEM signal is maintained.     

Lasso约束下融和光流信息的DCF目标跟踪算法

视频目标跟踪是计算机视觉的基础问题之一。近来由于 discriminative correlation filter（DCF）跟踪器的高效性和鲁棒性,出现了许多基于DCF的目标跟踪算法。为了克服DCF跟踪器对运动模糊目标的不适应性,本文提出了一种利用Lasso约束并融入光流信息的目标跟踪算法。首先在跟踪器抽取特征通道块中融入光流特征。然后在通道块之后进行多特征融合。其次利用Lasso约束DCF跟踪器的目标函数。考虑到所约束的目标函数在定义域上不连续和目标跟踪的优化效率。最后,采用块坐标下降算法来优化所约束的目标函数。实验结果表明,与基于DCF视觉跟踪算法相比,所提出的算法可以有效的处理运动模糊目标,实现复杂环境下鲁棒的视觉目标跟踪。      

基于改进概率数据关联滤波的红外小运动目标跟踪

在跟踪弱小运动目标过程中,低信噪比和高虚警率使得即使是在波门内也存在较多的虚假目标,加大目标识别的不确定性,降低跟踪的可靠性和精度。该文分析了红外成像系统中目标信号特点,根据概率数据关联滤波的思想,提出了一种基于目标幅度连续性和运动一致性的弱小目标跟踪识别方法,重点比较了极大似然-概率数据关联滤波和改进算法的目标识别与跟踪性能。理论分析和实验结果表明,该方法较大程度上减少了弱小目标跟踪过程中测量点与航迹关联的不确定性,提高了目标识别与跟踪性能。     

昆虫目标雷达散射截面积特征辅助跟踪算法

害虫迁飞具有规模大、突发性强的特点,会导致病虫害异地大爆发,粮食产量下降,造成重大的经济损失。昆虫雷达是监测迁飞性害虫的一种有效手段。昆虫目标的雷达散射截面积(RCS)较小,回波能量弱,在保证高检测率的同时会带来高虚警率问题,进而导致在目标跟踪的数据关联环节,易受虚假量测的影响出现关联错误。幅度特征辅助跟踪算法利用目标与噪声点迹的幅度差异,可以有效提高目标与噪声的识别度,改善跟踪性能,但是其需要已知目标的RCS起伏模型作为先验信息来计算幅度似然比。因此,该文基于Ku波段高分辨昆虫雷达外场实测昆虫回波数据,分析了昆虫目标的RCS起伏特性,得出Gamma分布可以较好地拟合昆虫目标的RCS统计分布,并将其作为先验信息,推导出Gamma起伏目标在高斯白噪声背景下的幅度似然比。通过在不同信噪比、不同量测噪声及不同起伏模型参数下的仿真结果及性能指标分析,验证了相比于概率数据互联滤波算法(PDAF)算法,目标RCS特征辅助的跟踪算法可以有效提高昆虫目标的跟踪精度。      

一种航迹恒虚警的目标检测跟踪一体化算法

对于传统雷达系统,目标检测和目标跟踪通常被看作是两个独立的过程。然而,在跟踪阶段,可以获得目标位置信息(跟踪信息),将该信息反馈至目标检测器将有助于提升检测性能。为此,该文提出一种航迹恒虚警的目标检测跟踪一体化算法。首先根据目标跟踪信息和目标运动模型,建立预测波门;然后按照给定的航迹虚警概率,计算帧虚警概率,即预测波门内至少出现一个虚警的概率;最终根据帧虚警概率调整预测波门内的检测门限,完成目标检测过程。仿真实验表明,该算法可以大大提高目标检测概率,推远目标跟踪距离,同时也可以保证目标突然消失时航迹能够以较高的概率结束。     

基于改进YOLO网络的双通道显著性目标识别算法

由于缺乏目标的先验信息,实时预警检测系统存在虚警率高、实时性偏低等问题,限制了实战环境下的广泛应用。为了提升目标检测识别的性能,本文提出了一种基于改进YOLO网络的双通道显著性目标识别算法,该算法利用红外图像与可见光互补特性进行多尺度融合,并在融合图像上采用显著性检测获取疑似目标区域,最后利用改进的识别网络对疑似区域进行多层次目标识别。改进的YOLO识别网络增加了一路辅助网络,改善整个特征提取网络的性能,并采用注意机制对辅助网络和骨干网络的特征信息融合,增强有效信息通道,抑制无效信息通道,提高网络识别效率。仿真实验结果表明,本文提出的模型可以有效地提高目标检测与识别精度,其实时性得到了大大增强。