Research on Discrimination Algorithm of Sandblasting Derusting Quality Based on Improved Texture Features.

Autor: Junyou Zhao, Liu, Zhongping, Yan, Chengxin, Dong, Yafei, Kwabena, A. R.
Zdroj: Automatic Control & Computer Sciences; Dec2021 Supplement 1, Vol. 55 Issue 1, p89-98, 10p
Abstrakt: With the development of smart oil-fields, intelligent sandblasting and rust removal is in urgent need of rust removal quality identification technology in the petroleum and petrochemical industry. Traditional algorithm for quality assessment of sandblasting rust removal has a single feature and a low recognition rate, and is based on the gray level co-occurrence matrix (GLCM) feature, utilizing the template matching (TM) classification algorithm. In this paper, an improved rust removal quality discrimination algorithm, which is a combination of GLCM features and local binary patterns (LBP), is proposed. This algorithm takes both the local and global texture features into account. LBP have advantage in rotation invariance. In addition, the proposed algorithm improved the recognition accuracy by using support vector machine (SVM) for learning and classification. Comparative experimentation shows that the improved algorithm has stronger feature robustness and better recognition rate than the traditional algorithm. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index
Externí odkaz: