塑壳地雷的声波探测与智能识别技术研究

In modern warfare, the metallic landmines are replaced gradually by the plastic landmines, which results in the unsatisfied performance of the traditional mine detector based on the electromagnetic induction method. Up to now, the detection of plastic landmines is an international problem to be solved. Based on the superior acoustic compliance of the plastic landmines and the coupling principle between the acoustic waves and seismic waves, the mine detection technology with acoustic excitation method reveals a promising application prospect in the aspects of mine safety detection. However, the anti-resonance effect, weakening the vibration amplitude of the surface, makes the current feature parameters is not enough to reflect the mechanical properties of the plastic landmine. Besides, the point by point and contact-type measurement method to the weak vibration of the ground surface limits the practical process of the detection system. Based on the lumped parameter linear model of the acoustic landmine detection system, the mechanism of the resonance, anti-resonance and multi-modal resonance will be explored according to the transfer function in the frequency domain. Based on the distributed nonlinear simulation model, the vibration characteristics will be investigated in the time domain, and the extraction method of characteristic parameters to unify the resonance and anti-resonance mechanism will be studied. Finally, the multi-level algorithms fusion method of the change detection, circle similarity matching and neural network will be adopted to improve the accuracy rate of the target recognition.

塑壳地雷大量出现并逐渐取代金属外壳地雷，使得传统的金属探雷器难以满足实际需求，塑壳地雷的探测成为一个公认的世界性难题。基于地雷声顺较大的机械特性和声波-地震波耦合原理的谐振技术在塑壳地雷的安全可靠探测方面有广阔的应用前景，但对地表振幅起削弱作用的反共振现象使得当前的特征参数不足以体现地雷的机械特性，对地表微弱振动逐点接触式的测量方法限制了声震探雷系统的实用化进程。本项目基于声震探雷的集中参数线性模型，利用传递函数在频率域开展塑料地雷的共振、反共振和多模态共振机理研究；基于声震探雷的分布式非线性仿真模型，在时域研究振动信号的特点并提出统一共振、反共振特征参数的表征方法；最后，研究变化检测、圆周相似度匹配、神经网络等算法的多层次融合方法，提高地雷目标识别的准确率。

塑壳地雷大量出现并逐渐取代金属外壳地雷，使得传统的金属探雷器难以满足实际需求，塑壳地雷的探测成为一个公认的世界性难题，基于地雷声顺较大的机械特性和声波-地震波耦合原理的谐振技术在塑壳地雷的安全可靠探测方面有广阔的应用前景。本课题针对声震探雷机理不够清晰、目标识别自动化程度低等问题，综合运用理论分析、试验研究、有限元仿真相结合的研究手段，建立了声-地震耦合探雷探测识别的集中参数线性模型和分布式仿真模型；通过样本数据分析，探索了声波激励下浅层土壤和地雷的振动特性，提出了不同情况下振动特征的利用方法；采用参量阵声源和自混合干涉仪为核心的激励-探测手段，构建了试验系统，实现了地表微米级微弱振动信号的实时检测；从时间域和空间域两个维度依次设计的信号滤波、特征提取、目标识别和多层次融合算法，能自动、快速、准确得到待测区域地雷数量、位置、尺寸。该课题研究成果可为地雷排扫、机动路径规划提供详细的信息支撑，可为声震探雷实用化系统的研制提供理论参考和技术支撑。