一种基于SVM的中密度纤维板表面粗糙缺陷检测方法  

An MDF Surface Roughness Detection Method Based on SVM

作　　者：宋辉[1] 刘佳玮 SONG Hui;LIU Jia-wei(Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学,辽宁沈阳110870

出　　处：《电脑知识与技术》2020年第13期36-38,共3页Computer Knowledge and Technology

摘　　要：中密度纤维板是目前家具制作市场上的重要原材料,其产品的美观性是决定其商业价值的重要因素之一。目前针对中密度纤维板外观上的表面粗糙缺陷通常采用人工检测方法,为实现板面粗糙的自动检测,本文提出一种基于SVM的中密度纤维板表面粗糙检测方法。采用线阵相机搭配远心镜头完成样品图像的采集,计算图像灰度均值与标准差,利用灰度共生矩阵提取能量、逆差分矩等5个特征,构建SVM进行训练、识别。训练图库70张,测试图库50张,该方法的识别准确率为96%。Medium density fiberboard(MDF)is an important raw material in the market of furniture making.At present,the surface roughness of MDF is usually detected by manual method.In order to automatically detect the surface roughness of MDF,a svmbased surface roughness detection method is proposed.The sample image was collected with a linear array camera and a telecenten⁃nial lens,the grayscale mean and standard deviation of the image were calculated,and the grayscale co-incidence matrix was used to extract five features,such as energy and deficit moment,and SVM was constructed for training and recognition.Seventy images were trained and fifty were tested,and the recognition accuracy of the method was 96%.

关 键 词：中密度纤维板 表面粗糙 纹理检测 灰度共生矩阵 SVM 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]