胡传双,邢益显,云虹,朱新波,李重根. 基于色彩识别特征的柳杉锯材表面活节和死节的机器视觉识别系统.[J]. 华南农业大学学报, 2009, 30(1). DOI: 10.7671/j.issn.1001-411X.2009.01.025
    引用本文: 胡传双,邢益显,云虹,朱新波,李重根. 基于色彩识别特征的柳杉锯材表面活节和死节的机器视觉识别系统.[J]. 华南农业大学学报, 2009, 30(1). DOI: 10.7671/j.issn.1001-411X.2009.01.025
    HU Chuan-shuang,XING Yi-xian,YUN Hong,ZHU Xin-bo,LI Chong-gen. Locating and Identifying Sound Knots and Dead Knots on Sugi by the Rule-Based Color Vision System[J]. Journal of South China Agricultural University, 2009, 30(1). DOI: 10.7671/j.issn.1001-411X.2009.01.025
    Citation: HU Chuan-shuang,XING Yi-xian,YUN Hong,ZHU Xin-bo,LI Chong-gen. Locating and Identifying Sound Knots and Dead Knots on Sugi by the Rule-Based Color Vision System[J]. Journal of South China Agricultural University, 2009, 30(1). DOI: 10.7671/j.issn.1001-411X.2009.01.025

    基于色彩识别特征的柳杉锯材表面活节和死节的机器视觉识别系统.

    Locating and Identifying Sound Knots and Dead Knots on Sugi by the Rule-Based Color Vision System

    • 摘要: 利用色彩特征信息开发了柳杉锯材表面活节和死节的机器视觉自动识别系统.该系统由3部分组成:CCD工业摄像图像采集硬件系统、缺陷检出的图像处理模块和基于识别规则的缺陷识别模块.潜在缺陷区域可由大津自动阈值分割算法结合T-检验来完成,活节和死节的检出率分别为92.6%和97.1%.基于2个形状识别特征和6个色彩识别特征构建了缺陷的识别规则,利用构建的识别规则可实现活节和死节的识别率分别为92.0%和94.1%.系统整体检测准确率为87.6%,此结果表明基于识别规则的彩色机器视觉自动识别系统是检测柳杉锯材表面活节和死节的一个有效手段.

       

      Abstract: The split and the hole are two common defects on sugi,Cryptomeria japonica.They have a common feature in that they are associated with surface irregularities.A laser scanning system to detect the splits and the holes based on their thickness was developed,which correlated spatially with the profile information.The displacements measured by the laser sensor were converted to pixel values to generate the displacement profile image.Both the splits and the holes manifested well in the image.A dedicated image-processing program written in Visual Basic was developed.The defects regions were accurately located by the image processing.To identify the defects,eight recognition rules based on four features were utilized.Furthermore,a method based on the pixel model was proposed to compute the area of the defect.The results indicated that the defects could be identified correctly,and the areas could be computed accurately using the pixels model.

       

    /

    返回文章
    返回