基于离散Gabor变换的磁暴识别

Automatic recognition of geomagnetic storm based on discrete Gabor Transform

以地磁秒数据为研究对象,通过离散Gabor变换将时域的地磁数据转换至二维时频面,提取Gabor变换谱图的均值和方差作为特征值,使用支持向量机实现地磁正常数据与磁暴干扰数据的自动分类识别。对5个地磁台的200组地磁秒数据进行计算分析,结果表明该方法对测试样本数据的识别率可达94%。

Taking the geomagnetic data with a sample rate of 1 as a research object,we propose an approach to extract eigenvalues of geomagnetic data based on discrete Gabor Transform,and then adopt the support vector machine to classify normal data and geomagnetic storm interference data.The experiment results indicate that the recognition rate is above 94%for 200 sets of geomagnetic data.