Evaluation of surface roughness in incremental forming using image processing based methods

•Explored different image processing methods for surface roughness evaluation in ISF.•Compared the efficiency of Euclidian, Hamming and Wavelet based methods.•The wavelet based method found to be more efficient over other methods.•The work can be extended to in-process measurement of roughness in re...

Full description

Saved in:
Bibliographic Details
Published in:Measurement : journal of the International Measurement Confederation Vol. 164; p. 108055
Main Authors: Gandla, Praveen Kumar, Inturi, Vamsi, Kurra, Suresh, Radhika, Sudha
Format: Journal Article
Language:English
Published: London Elsevier Ltd 01.11.2020
Elsevier Science Ltd
Subjects:
Classification
Euclidian distance
Evaluation
Factorial design
Hamming distance
Image processing
Image processing systems
Incremental sheet forming
On–line measurement
Process parameters
Studies
Surface roughness
Three dimensional imaging
Wavelet analysis
Wavelet transform
Wavelet transforms
Incremental sheet forming
On–line measurement
Surface roughness
Wavelet transform
Hamming distance
Euclidian distance
ISSN:0263-2241, 1873-412X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Explored different image processing methods for surface roughness evaluation in ISF.•Compared the efficiency of Euclidian, Hamming and Wavelet based methods.•The wavelet based method found to be more efficient over other methods.•The work can be extended to in-process measurement of roughness in real time. The present paper focuses on evaluation of surface roughness (Ra) in incrementally formed parts by different image processing based methods. For this, twenty-seven parts are formed as per full factorial design in incremental forming by varying three important process parameters over three levels each. The surface roughness of formed parts is measured using Taylor Hobson Talysurf and it is in the range of 0.6–3.6 µm. For image processing based evaluation of surface roughness, five images are captured from each formed part and created an image database. These images are classified in to three different classes based on the range of surface roughness using Euclidian distance method, Hamming distance method and Wavelet based method. The results reveal that the wavelet based method has yielded highest classification efficiency of 95.4%. The Hamming and Euclidian distance methods have a classification efficiency of 78.39% and 81.48% respectively.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0263-2241
1873-412X
DOI:10.1016/j.measurement.2020.108055