首页|基于主成分分析的随机森林钢材缺陷检测算法

基于主成分分析的随机森林钢材缺陷检测算法

扫码查看
钢材在原材料生产、冶炼加工等过程中会出现划痕、污渍等缺陷问题,而缺陷检测是生产过程中把控产品质量和生产效率不可或缺的一环.传统的分类算法对缺陷检测来说效率较低,因此本文提出了一种基于主成分分析和SMOTE的随机森林算法.通过融合SMOTE、主成分分析和随机森林算法,实现对钢材缺陷的有效检测.结果表明,相较于传统分类算法如Logistic回归、决策树等,本文提出的方法有效提升了缺陷检测效率.
Random forest steel defect detection algorithm based on PCA
Steel materials can have defects such as scratches and stains during the raw material production,smelting,and processing processes.Defect detection is an essential aspect of controlling product quality and production efficiency in the manufacturing process.Traditional classification algorithms are inefficient for defect detection.Therefore,a random forest algorithm based on PCA and SMOTE is proposed in this work.Through SMOTE,PCA,and random forest algorithm,effective detection of steel defects has been achieved.Experimental results demonstrate that compared to traditional classification algorithms such as Logistic Regression and Decision Trees,the proposed method significantly improves defect detection efficiency.

defect detectionSMOTEPCArandom forest algorithm

王纯杰、张钺、谭佳伟

展开 >

长春工业大学 数学与统计学院,吉林 长春 130012

缺陷检测 SMOTE 主成分分析 随机森林算法

国家自然科学基金项目国家自然科学基金项目吉林省科技厅重大科技专项项目吉林大学符号计算与知识工程教育部重点实验室项目

122264161227106020210301038GX93K172021K10

2024

吉林师范大学学报(自然科学版)
吉林师范大学

吉林师范大学学报(自然科学版)

影响因子:0.397
ISSN:1674-3873
年,卷(期):2024.45(1)
  • 7