摘要
研究混凝土的配合比设计是混凝土工程中的一项重要工作.为了实现不同粒径大小以及砼、砖配合比的混凝土图像的精确分类,提出了一种基于彩色共生矩阵(Color GLCM,CGLCM)和支持向量机(Support Vector Machine,SVM)的混凝土图像分类方法.首先,将混凝土图像从RGB空间转换到HSV空间,分别提取H,S,V 3个通道的8维灰度共生矩阵(Gray level co-occurrence matrix,GLCM)的纹理特征;其次,将3个通道的特征进行融合,构成24维特征向量,并将特征向量输入到SVM进行分类,确定CGLCM的最佳参数;最后,在此基础上利用群智能算法对SVM参数进行优化.通过对中北大学土木工程专业构建的两种数据集(CIRD_A,CIRD_B)进行分类,结果表明,与其他分类方法相比,遗传算法优化SVM模型(GA-SVM)对两种数据集的分类精度最好,分别达到了97.76%和96.34%.
It is an important work to study the mix proportion design of concrete in concrete engineering.In order to realize the accurate classification of concrete with different particle size and mix proportion of concrete and brick,a concrete image classification method based on Color GLCM(CGLCM)and Support Vector Machine(SVM)was proposed.Firstly,the concrete image was transformed from RGB space to HSV space,and the 8-dimensional Gray Level Co-occurrence Matrix(GLCM)texture features of H,S and V channels were extracted respectively.The features of the three channels were fused to form a 24 dimensional feature vector.The feature vectors were input into SVM for classification,and the best parameters of CGLCM were determined by experiments.On this basis,Swarm intelligence algorithm was used to optimize the parameters of SVM.The classification experiment was carried out with two kinds of data sets(CIRD_A,CIRD_B)constructed by civil engineering specialty of North University of China.The experimental results show that compared with other classification methods,GA-SVM has the best classification accuracy for the two kinds of datasets,reaching 97.76%and 96.34%respectively.
作者
张莉
焦宇倩
续婷
侯宇超
白艳萍
李建军
ZHANG Li;JIAO Yuqian;XU Ting;HOU Yuchao;BAI Yanping;LI Jianjun(School of Science, North University of China, Taiyuan 030051, China;School of Information and Communication Engineering, North University of China, Taiyuan 030051, China)
出处
《中北大学学报（自然科学版）》
CAS
2022年第1期84-90,共7页
Journal of North University of China(Natural Science Edition)
基金
国家自然科学基金资助项目(61774137)
山西省重点研发计划项目(201903D121156)
山西省自然科学基金资助项目(201801D121026)。
