基于大容量棉花检测仪检测结果的进口锯齿棉品级分类  

Grade classification of imported sawtooth cotton based on HVI test results

作　　者：董绍伟 徐静[2] 刘俊 刘冬志 DONG Shaowei;XU Jing;LIU Jun;LIU Dongzhi(Nanjing Customs District Industrial Products Inspection Center,Nanjing,Jiangsu 210019,China;Basic Department of Yangzhou Polytechnic Institute,Yangzhou,Jiangsu 225127,China;Technology Center of Chengdu Customs,Chengdu,Sichuan 610041,China;Huanjiang Lüyuan Environmental Protection Technology Co.,Ltd.,Urumqi,Xinjiang 831400,China)

机构地区：[1]南京海关工业产品检测中心,江苏南京210019 [2]扬州工业职业技术学院基础部,江苏扬州225127 [3]成都海关技术中心,四川成都610041 [4]新疆环疆绿源环保科技有限公司,新疆乌鲁木齐831400

出　　处：《毛纺科技》2023年第4期82-86,共5页Wool Textile Journal

基　　金：新疆维吾尔自治区重大科技专项项目(2020A03002-4)。

摘　　要：为了消除进口棉花品级分类中由于检测人员主观性引起的观察偏倚,以及其与预警信息化和及时性不相符等矛盾,文章基于大容量棉花检测仪(HVI)的检测结果,提出运用随机森林算法来预测进口锯齿棉品级分类的方法。对HVI检测获得的原始数据进行标准化处理,并运用主成分分析降低指标维度,采用随机森林算法构建分类模型,将测试样本带入训练好的分类模型,可自动判别出测试锯齿棉的品级。实验结果表明该方法的分类效果较好,可有效实现锯齿棉品级的自动分类。Based on the HVI test results,a random forest algorithm was used to predict the grade classification of imported sawtooth cotton.This method can eliminate the observational bias caused by subjectivity in the classification of imported cotton grades,and remove contradictions that are inconsistent with the informatization and timeliness of early warning.The original data obtained by HVI were standardized,the index dimension was reduced by using principal component analysis,and the classification model was constructed by using random forest algorithm.The test samples were brought into the trained classification model,and the grades of the tested sawtooth cotton could be automatically discriminated.The experimental results show that the classification effect of this method is good,and it can effectively realize the automatic classification of sawtooth cotton grades.

关 键 词：HVI检测 锯齿棉 随机森林算法 品级分类 

分 类 号：TS101.9[轻工技术与工程—纺织工程] S562[轻工技术与工程—纺织科学与工程]