论文部分内容阅读
为了改进基于震动信号的地面运动目标识别算法,提出了一种基于主成分分析(PCA)的2次特征提取算法.首先对地面运动目标引起的震动信号进行目标特性分析,提取多维的特征值;然后利用主成分分析方法对众多的特征值进行分析,去除特征值之间的相关性,提取综合特征值并应用于分类器,得到目标识别结果.基于实地采集的地面运动目标的震动信号进行实验,结果表明:该方法有效地减少了特征值的维数和相关性,降低了分类器训练的难度和训练时间,同时提高了目标的正确识别率.
In order to improve the ground moving target recognition algorithm based on the vibration signal, this paper proposes a PCA - based two - time feature extraction algorithm, which firstly analyzes the target characteristics of ground motion caused by vibration signals and extracts multi - dimensional eigenvalues. Then, the principal component analysis method is used to analyze many eigenvalues, remove the correlation between the eigenvalues, extract the integrated eigenvalues and apply them to the classifier to get the target identification results.According to the vibration signal of the ground moving target collected in the field, The results show that this method effectively reduces the dimensionality and correlation of eigenvalues, reduces the difficulty of classifier training and training time, and improves the correct recognition rate of the target.