TY - JOUR

T1 - A Wavelet Packet Tree Denoising Algorithm for Images of Atomic-Force Microscopy

AU - Schimmack, Manuel

AU - Mercorelli, Paolo

PY - 2018/7

Y1 - 2018/7

N2 - A threshold-free denoising procedure of acquired discrete Atomic-force microscopy (AFM) signals using the discrete wavelet transform (DWT) method is presented in this article. The integration of a denoising procedure into a control structure is extremely important for each kind of system to be controlled. The detection of unavoidable measurement noise in the acquired data of the AFM signal is done by using orthogonal wavelets (Daubechies and Symmlet) and with different polynomial approximation order for each family. The proposed denoising algorithm, based on the free wavelet toolboxes from the WaveLab 850 library of the Stanford University (USA), compares the usefulness of Daubechies and Symmlet wavelets with different vanishing moments. With the help of a seminorm the noise of a sequence is defined as a coherent and incoherent part of the AFM signal. In the first step of the procedure the algorithm analyzes the frequency subspaces of the wavelet packets tree and searches for small or opposing components in the wavelet domains. In the second step of the procedure the incoherent components in the low- and high frequency domains are localized and the incoherent is then removed from the AFM signal. The proposed algorithm structure is used to improve the quality of the AFM signals and it can be easily integrated into the existing AFM control hard- and software structures. The effectiveness of the proposed denoising algorithm is validated with real measurements.

AB - A threshold-free denoising procedure of acquired discrete Atomic-force microscopy (AFM) signals using the discrete wavelet transform (DWT) method is presented in this article. The integration of a denoising procedure into a control structure is extremely important for each kind of system to be controlled. The detection of unavoidable measurement noise in the acquired data of the AFM signal is done by using orthogonal wavelets (Daubechies and Symmlet) and with different polynomial approximation order for each family. The proposed denoising algorithm, based on the free wavelet toolboxes from the WaveLab 850 library of the Stanford University (USA), compares the usefulness of Daubechies and Symmlet wavelets with different vanishing moments. With the help of a seminorm the noise of a sequence is defined as a coherent and incoherent part of the AFM signal. In the first step of the procedure the algorithm analyzes the frequency subspaces of the wavelet packets tree and searches for small or opposing components in the wavelet domains. In the second step of the procedure the incoherent components in the low- and high frequency domains are localized and the incoherent is then removed from the AFM signal. The proposed algorithm structure is used to improve the quality of the AFM signals and it can be easily integrated into the existing AFM control hard- and software structures. The effectiveness of the proposed denoising algorithm is validated with real measurements.

KW - Atomic-force microscopy

KW - Denoising algorithm

KW - Orthogonal wavelets

KW - Wavelet packets transform method

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=85040168585&partnerID=8YFLogxK

U2 - 10.1002/asjc.1718

DO - 10.1002/asjc.1718

M3 - Journal articles

AN - SCOPUS:85040168585

VL - 20

SP - 1367

EP - 1378

JO - Asian Journal of Control

JF - Asian Journal of Control

SN - 1561-8625

IS - 4

ER -