【作者】 孙磊；

【导师】 柏逢明；

【作者基本信息】 长春理工大学 ， 模式识别与智能系统， 2011， 硕士

【摘要】 在现代军事训练或者竞技体育比赛中,射击训练的结果检测是必不可少的一个重要环节。目前,我国传统的以人工操作为基础的报靶方法存在着严重的问题,例如报靶精度差、效率低、存在人力资源的浪费,并且在报靶的过程中有一定的安全风险。为了提高射击训练水平和安全系数的提高,对报靶系统的研究需求很强烈。本文分析了激波和声爆理论,构建出超声智能报靶系统的总体设计方案,分析了超声波传感器阵列的布阵方法,对弹着点坐标计算的数学公式进行了研究和推导,然后研究了弹丸激波信号的小波阈值去噪处理,去噪后的信号送入处理器进行信息处理,最后通过上位机的软件编程将报靶结果显示出来。本论文的主要工作是弹丸激波信号的去噪处理和传感器阵列分布的研究。小波去噪处理的仿真实验表明去噪算法的选择具有提高系统信噪比的特点,而且它降低了均方根误差。在传感器的阵列分布研究中,提出了十一点阵阵列定位模型,并且通过遗传神经网络算法对数学模型中的加权系数进行了优化选择处理,进一步的降低了报靶系统的误差。更多还原

【Abstract】 In the modern military training or sports competition, the results of test in firing training is an important part. At present, traditional target reporting which uses manual-based method exists some serious problems, for example, poor accuracy, low efficiency, waste of human resources and a certain security risk. In order to improve the level of training and safety factor, the research of automatic target system is strongly needed.This paper analyzes the theory of shock wave and sonic boom, proposes the composing figure of the intelligent target-scoring system, analyzes the method of ultrasonic sensor array, discusses mathematical formula of point coordinate of bullet, then rearches the wavelet thresholding denoising of bullet shock wave signal, transmits the denoising signal to the processor and processes the information,reports the results of target-scoring through software programming of the PC. The main content of this thesis contains denoising of bullet shock wave and the distribution of sensor arrays. The simulation results of wavelet thresholding denoising show that the choice of denoising algorithm can improve signal noise ratio of the system and reduce the mean square error. On researching the distribution of sensor array, this paper proposes the location model of 11 lattices, and optimizes the weighting coefficient by genetic neural network algorithm, further reduces the error of the target scoring system.更多还原