Open AccessApplication of update lifting morphological wavelet and non-negative matrix factorization for wheeled and tracked vehicles classification
Author(s) -
Kai Ding,
Kai Du,
Xiaogang Qi,
Yuelin Xu,
Jingshi Cui
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/660/1/012083
Subject(s) - wavelet , lifting scheme , non negative matrix factorization , computer science , filter (signal processing) , matrix decomposition , pattern recognition (psychology) , signal (programming language) , artificial intelligence , scheme (mathematics) , matrix (chemical analysis) , wavelet transform , operator (biology) , computation , algorithm , discrete wavelet transform , computer vision , mathematics , eigenvalues and eigenvectors , materials science , repressor , mathematical analysis , chemistry , composite material , biochemistry , quantum mechanics , transcription factor , programming language , physics , gene
This paper presents an available scheme based on update lifting morphological wavelet (ULMW) and Non-negative matrix factorization (NMF) for wheeled and tracked vehicles classification. The ULWM algorithm which utilizes the update operator, means the morphological filter to replace the linear filter can preserve the impulsive shape details in seismic signal. Meanwhile the NMF method can reduce the computation cost. The traditional linear wavelet analysis and statistical analysis are compared with the presented scheme. Experimental results demonstrate that the presented scheme achieves a promising performance on extracting impulsive features of seismic signal and recognizing ground moving target.