基于尾焰光谱分布特性分析的高速目标识别系统

发布时间：2018-10-31 16:26

【摘要】：为了实现对高速运动目标的快速识别,设计了一种基于尾焰光谱分布特征分析的识别系统。系统由跟踪成像模块和光谱分析模块组成,跟踪成像模块用于对准目标,光谱分析模块用于目标识别。在此基础上,研究了基于多特征波段峰均值与谷均值比例的区分因子识别算法。该算法设置了有效区分区间,并通过区分因子比例关系实现高速目标的识别。分析了目标高速运动对光谱获取的影响,给出了决定光谱偏移量大小的径向和切向速度的函数关系。实验采用少量火箭弹燃烧部作为被测目标,在距探测系统0.5 km、1.0 km、2.0 km以及4.0 km处分别以10.0 m/s~100.0 m/s的速度进行光谱探测实验。实验结果显示,对于不同速度的被测目标而言,在相等的采样周期内系统检测的光谱分布产生了明显的光谱偏移,但其光谱分布形态基本一致;对于不同测试距离而言,距离越远能量衰减越强,虽然光谱整体振幅差异很大,但光谱形态无明显变化。通过计算相应波段上峰均值和谷均值的区分因子与区分区间的包含关系,完成了对被测目标的有效识别。
[Abstract]:In order to realize fast recognition of high speed moving target, a recognition system based on spectral distribution analysis of tail flame is designed. The system is composed of tracking imaging module and spectral analysis module. The tracking imaging module is used for target alignment and the spectral analysis module is used for target recognition. On this basis, a discriminant factor recognition algorithm based on the ratio of multiband peak mean and valley mean is studied. The algorithm sets an effective discriminant interval, and realizes the recognition of high speed targets by the proportion relation of discriminant factors. The influence of high speed motion of target on spectral acquisition is analyzed, and the function relationship between radial and tangential velocities which determine the spectral offset is given. A small number of rocket combustion units were used as the measured targets, and the spectral detection experiments were carried out at the velocity of 10.0 m/s~100.0 m / s at 0. 5 km,1.0 km,2.0 km and 4. 0 km from the detection system, respectively. The experimental results show that, for the target with different velocities, the spectral distribution detected by the system has obvious spectral offset in the same sampling period, but the spectral distribution is basically the same. For different measuring distances, the longer the distance is, the stronger the energy decay is. Although the amplitude of the spectrum varies greatly, the spectral morphology has no obvious change. The effective recognition of the target under detection is accomplished by calculating the inclusion relation between the distinction factor of the peak mean and the valley mean in the corresponding band and the discriminant interval.
【作者单位】：长春理工大学光电信息学院;
【基金】：总装国防科技基金国家自然基金(51305409)
【分类号】：O433.4

【参考文献】