摘要
在客观条件下,获取高光谱数据会产生大量噪声,最终会影响数据分析,传统的小波滤波由于基函数的选择从而缺乏自适应性。文章基于EMD的算法特性,在考虑高光谱数据的变化特征的基础上,采用了一种基于连续均方差准则的EMD去噪方法。结果表明:数据在EMD分解后,噪声与信号能够有效地分离,具有对高光谱数据进行滤波的可行性。根据实例,此法滤波后信噪比为44.84,而小波滤波信噪比为42.78,可见此方法优于小波滤波。
Under objective conditions,acquiring hyperspectral data will generate a lot of noise,which will eventually affect data analysis. However, the traditional wavelet filter lacks self-adaptability in denoising due to the selection of the basis function. Based on the algorithm characteristics of EMD,this paper adopts an EMD denoising method based on the continuous mean square error criterion based on changing characteristics of hyperspectral data. The results show that noise and signal can be separated effectively after EMD decomposition of data,and it is feasible to filter hyperspectral data. According to the analysis of an example,the method reduces the signal-to-noise ratio to 44.84 while the signal-to-noise ratio of wavelet filtering is 42.78. It can be seen that this method is better than wavelet filtering.
作者
武进
Wu Jin(Key Laboratory of Cloud Computing and Intelligent Information Processing ,Sichuan Technology and Business University ,Chengdu 611745 China)
出处
《四川工商学院学术新视野》
2019年第1期24-28,共5页
Academic New Vision of Sichuan Technology and Business University
关键词
连续均方差
EMD
高光谱数据
滤波
Continuous mean square error
EMD
Hyperspectral data
Filtering
