Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm

Considering the tiny surface defects of Si3N4 ceramic bearing balls and the low accuracy of defect detection using a single traditional algorithm, we find that the performance of the aerospace mechanical power system is poor. A coupling algorithm based on the improved homomorphic filter and Gaussian...

Full description

Bibliographic Details
Main Authors: Dahai Liao, Zhihui Cui, Jun Li, Wenjie Li, Wei Wang
Format: Article
Language:English
Published: AIP Publishing LLC 2022-02-01
Series:AIP Advances
Online Access:http://dx.doi.org/10.1063/5.0082702
_version_ 1818488142507802624
author Dahai Liao
Zhihui Cui
Jun Li
Wenjie Li
Wei Wang
author_facet Dahai Liao
Zhihui Cui
Jun Li
Wenjie Li
Wei Wang
author_sort Dahai Liao
collection DOAJ
description Considering the tiny surface defects of Si3N4 ceramic bearing balls and the low accuracy of defect detection using a single traditional algorithm, we find that the performance of the aerospace mechanical power system is poor. A coupling algorithm based on the improved homomorphic filter and Gaussian filter is proposed. The Si3N4 ceramic bearing ball nondestructive testing platform is established, by which the surface defect images of Si3N4 ceramic bearing balls are collected. The image acquisition card obtains the image information and then transmits it to the image storage module. The image processing module handles surface defect images. The gray transformation algorithm is used to complete image preprocessing. Gaussian noise in images is filtered out using the Gaussian filter algorithm. The homomorphic filtering algorithm is used to enhance the high frequency component, compress the low frequency component, and filter out the convolution noise and promiscuous signal. The contrast of the defect part has been reinforced using the coupling algorithm. It turns out that the accuracy of the coupling algorithm is 100%, 96.7%, 98.9%, and 94.4%.
first_indexed 2024-12-10T16:47:18Z
format Article
id doaj.art-7cffa8c43b93423c91f2baf9e5c48c27
institution Directory Open Access Journal
issn 2158-3226
language English
last_indexed 2024-12-10T16:47:18Z
publishDate 2022-02-01
publisher AIP Publishing LLC
record_format Article
series AIP Advances
spelling doaj.art-7cffa8c43b93423c91f2baf9e5c48c272022-12-22T01:41:01ZengAIP Publishing LLCAIP Advances2158-32262022-02-01122025325025325-1010.1063/5.0082702Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithmDahai Liao0Zhihui Cui1Jun Li2Wenjie Li3Wei Wang4School of Mechanical and Electronic Engineering, Jingdezhen Ceramic Institute, Jingdezhen, People’s Republic of ChinaSchool of Mechanical and Electronic Engineering, Jingdezhen Ceramic Institute, Jingdezhen, People’s Republic of ChinaSchool of Mechanical and Electronic Engineering, Jingdezhen Ceramic Institute, Jingdezhen, People’s Republic of ChinaSchool of Mechanical and Electronic Engineering, Jingdezhen Ceramic Institute, Jingdezhen, People’s Republic of ChinaSchool of Mechanical and Electronic Engineering, Jingdezhen Ceramic Institute, Jingdezhen, People’s Republic of ChinaConsidering the tiny surface defects of Si3N4 ceramic bearing balls and the low accuracy of defect detection using a single traditional algorithm, we find that the performance of the aerospace mechanical power system is poor. A coupling algorithm based on the improved homomorphic filter and Gaussian filter is proposed. The Si3N4 ceramic bearing ball nondestructive testing platform is established, by which the surface defect images of Si3N4 ceramic bearing balls are collected. The image acquisition card obtains the image information and then transmits it to the image storage module. The image processing module handles surface defect images. The gray transformation algorithm is used to complete image preprocessing. Gaussian noise in images is filtered out using the Gaussian filter algorithm. The homomorphic filtering algorithm is used to enhance the high frequency component, compress the low frequency component, and filter out the convolution noise and promiscuous signal. The contrast of the defect part has been reinforced using the coupling algorithm. It turns out that the accuracy of the coupling algorithm is 100%, 96.7%, 98.9%, and 94.4%.http://dx.doi.org/10.1063/5.0082702
spellingShingle Dahai Liao
Zhihui Cui
Jun Li
Wenjie Li
Wei Wang
Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
AIP Advances
title Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
title_full Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
title_fullStr Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
title_full_unstemmed Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
title_short Surface defect detection of Si3N4 ceramic bearing ball based on improved homomorphic filter-Gaussian filter coupling algorithm
title_sort surface defect detection of si3n4 ceramic bearing ball based on improved homomorphic filter gaussian filter coupling algorithm
url http://dx.doi.org/10.1063/5.0082702
work_keys_str_mv AT dahailiao surfacedefectdetectionofsi3n4ceramicbearingballbasedonimprovedhomomorphicfiltergaussianfiltercouplingalgorithm
AT zhihuicui surfacedefectdetectionofsi3n4ceramicbearingballbasedonimprovedhomomorphicfiltergaussianfiltercouplingalgorithm
AT junli surfacedefectdetectionofsi3n4ceramicbearingballbasedonimprovedhomomorphicfiltergaussianfiltercouplingalgorithm
AT wenjieli surfacedefectdetectionofsi3n4ceramicbearingballbasedonimprovedhomomorphicfiltergaussianfiltercouplingalgorithm
AT weiwang surfacedefectdetectionofsi3n4ceramicbearingballbasedonimprovedhomomorphicfiltergaussianfiltercouplingalgorithm