基于EWT与GA-SVM的地面目标识别

Ground target recognition based on EWT and GA-SVM

为提高地面移动目标识别效率,提出基于EWT(经验小波变换)与GA-SVM(遗传优化支持向量机)的目标识别方法。对采集到的地震动信号进行小波包去噪,对去噪信号进行经验小波变换,提取信号特征并构建特征向量输入到支持向量机中训练;利用遗传算法参数寻优,找到支持向量机的最佳参数,以此构建预测模型;利用训练完成的GA-SVM模型对目标进行分类。实验结果表明,GA-SVM模型优于未改进的交叉验证法SVM模型,对常见地面移动目标具有很好的分类准确率。

To improve the accuracy of ground moving target recognition,a kind of method based on the empirical wavelet transform(EWT)and genetic algorithm optimization support vector machine(GA-SVM)for target recognition was presented.The noise was reduced via wavelet packet.The EWT was applied to extract the features from the seismic signal after reducing noise and feature vector was created as the input of the SVM for training.The genetic algorithm were employed to search the optimal output parameters of the SVM and the SVM prediction model was constructed.The trained GA-SVM model was used to recognize the targets.Experimental results show that the GA-SVM model is better than the unimproved cross validation SVM model and it has good classification accuracy for common ground moving target.