TY - GEN
T1 - Wear Particles Classification Using Shape Features
AU - Laghari, Mohammad Shakeel
AU - Hassan, Ahmed
AU - Noman, Mubashir
N1 - Publisher Copyright:
© 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2021
Y1 - 2021
N2 - Wear debris provides important information about the machine condition that can be used to prevent the loss of expensive machinery. This information is crucial as it portrays the condition of the machines and can be used to predict early failure of the machinery that can prevent a major loss to the industry. Wear debris or particles produced in different parts of machine vary in shape, size, color, and texture. These characteristic features can be utilized to identify the type of wear debris. Human experts are extremely efficient in recognizing such objects; however, wear judgments are occasionally based on their specific perceptions. The goal is to look beyond the personal opinions and bring consistency in judging and recognizing wear particles. Keeping in view the above findings, this study focuses on the identification of wear particles by using shape-based features only. Different shape features, which include the Histogram of Oriented Gradients (HOG), Rotation Scale Translation (RST) invariant features, solidity, aspect ratio, circularity, and Euler number, are extracted and used to train three classification models of Support Vector Machine (SVM), k-Nearest Neighbors (kNN), and discriminant analysis (DA) classifier. The performance of the classifiers are compared with each another and classification of debris based on shape features is analyzed.
AB - Wear debris provides important information about the machine condition that can be used to prevent the loss of expensive machinery. This information is crucial as it portrays the condition of the machines and can be used to predict early failure of the machinery that can prevent a major loss to the industry. Wear debris or particles produced in different parts of machine vary in shape, size, color, and texture. These characteristic features can be utilized to identify the type of wear debris. Human experts are extremely efficient in recognizing such objects; however, wear judgments are occasionally based on their specific perceptions. The goal is to look beyond the personal opinions and bring consistency in judging and recognizing wear particles. Keeping in view the above findings, this study focuses on the identification of wear particles by using shape-based features only. Different shape features, which include the Histogram of Oriented Gradients (HOG), Rotation Scale Translation (RST) invariant features, solidity, aspect ratio, circularity, and Euler number, are extracted and used to train three classification models of Support Vector Machine (SVM), k-Nearest Neighbors (kNN), and discriminant analysis (DA) classifier. The performance of the classifiers are compared with each another and classification of debris based on shape features is analyzed.
KW - Histogram of oriented gradients
KW - Particles classification
KW - RST invariant features
KW - Shape-based features
KW - Wear debris
UR - http://www.scopus.com/inward/record.url?scp=85097093807&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85097093807&partnerID=8YFLogxK
U2 - 10.1007/978-981-15-8603-3_33
DO - 10.1007/978-981-15-8603-3_33
M3 - Conference contribution
AN - SCOPUS:85097093807
SN - 9789811586026
T3 - Advances in Intelligent Systems and Computing
SP - 377
EP - 385
BT - Proceedings of 6th International Conference on Harmony Search, Soft Computing and Applications - ICHSA 2020
A2 - Nigdeli, Sinan Melih
A2 - Bekdas, Gebrail
A2 - Kim, Joong Hoon
A2 - Yadav, Anupam
PB - Springer Science and Business Media Deutschland GmbH
T2 - 6th International Conference on Harmony Search, Soft Computing and Applications, ICHSA 2020
Y2 - 22 April 2020 through 24 April 2020
ER -