基于图像法和粒度分形模型的打叶过程烟片结构在线检测

    Online detection of tobacco lamina structure based on fractal and image analysis method

    • 摘要: 为实现烟叶采收后初加工过程打叶质量的实时监测与调控,该文采用图像分析法在线检测烟叶打后叶片结构,即叶片面积分布,应用分形理论建立烟叶打后叶片面积分布的分形模型,用分维数表征叶片面积分布特征。结果表明:分形模式计算的打后叶片面积分布结果与实测值相关系数较高,粒度分形模式可较准确描述烟叶打后叶片面积分布;分维数可表征打后叶片面积分布均匀程度,可反映出不同等级烟叶打叶特性差异;分维数与目前打叶过程的主要工艺控制指标大中片率成线性关系,相关系数大于0.90。基于图像法和粒度分形模型建立的打叶过程烟叶结构在线检测方法,可用于打叶质量的实时监测与调控。

       

      Abstract: In order to monitor and control tobacco lamina quality in threshing, image analysis method was used to measure tobacco lamina area. The fractal model of area distribution of tobacco strips was established based on the fractal theory, and fractal dimension was used to describe the characteristic of area distribution of tobacco strips. The results showed that the model predictions and measured values of area distribution of tobacco strip have high correlation coefficient. Fractal model can describe accurately area distribution of tobacco strip. Fractal dimension could describe the uniformity of tobacco strip structure. For tobacco leafs of different class, the fractal dimension could reflect tobacco characteristic difference in threshing. There is a liner relation between fractal dimension and percentage of >161.29 mm2 lamina, and the correlation coefficient was greater than 0.90. The determining method for tobacco lamina structure based on image analysis and fractal model could monitor and control tobacco lamina quality on line during threshing process.

       

    /

    返回文章
    返回