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

Dahai Liao,Wei Wang,Zhihui Cui,Jun Li,Wenjie Li

doi:10.1063/5.0082702

Abstract

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%.

Highlights

When it comes to the unique excellent performance of Si3N4 ceramic bearings, it lies in high mechanical strength and hardness, good chemical stability, thermal shock resistance, wear resistance, and self-lubrication.1–3 As an important part of the rotating mechanism, Si3N4 ceramic bearings are widely used in aerospace machinery, precision instruments, and other fields.4 They have broad application prospects as well.5,6 Due to the hardness and brittleness, the surface of ceramic bearing balls is prone to produce defects in the preparation and processing of Si3N4 ceramic bearings
Combining the Si3N4 nondestructive testing experimental platform with an image processing module, a surface defect detection method of Si3N4 ceramic bearing balls based on the improved homomorphic filtering and the Gaussian filtering coupling algorithm is proposed
Gaussian filtering and homomorphic filtering are combined to a coupling denoising algorithm

Summary

Introduction

When it comes to the unique excellent performance of Si3N4 ceramic bearings, it lies in high mechanical strength and hardness, good chemical stability, thermal shock resistance, wear resistance, and self-lubrication. As an important part of the rotating mechanism, Si3N4 ceramic bearings are widely used in aerospace machinery, precision instruments, and other fields. They have broad application prospects as well. Due to the hardness and brittleness, the surface of ceramic bearing balls is prone to produce defects in the preparation and processing of Si3N4 ceramic bearings. As an important part of the rotating mechanism, Si3N4 ceramic bearings are widely used in aerospace machinery, precision instruments, and other fields.. As an important part of the rotating mechanism, Si3N4 ceramic bearings are widely used in aerospace machinery, precision instruments, and other fields.4 They have broad application prospects as well.. Due to the hardness and brittleness, the surface of ceramic bearing balls is prone to produce defects in the preparation and processing of Si3N4 ceramic bearings. These defects include pits, wears, scratches, snowflakes, and cracks.. At present, detecting the surface defects of Si3N4 bearings mainly depend on manual visual inspection.. Combining the Si3N4 nondestructive testing experimental platform with an image processing module, a surface defect detection method of Si3N4 ceramic bearing balls based on the improved homomorphic filtering and the Gaussian filtering coupling algorithm is proposed

Methods

Results

Conclusion