针对一般烟叶宽度提取算法准确性不高及不适应烟叶随机摆放的问题,提出一种基于旋转矢量不变性的烟叶宽度提取算法。该方法首先使用最小外接矩形法获得烟叶的首尾端点,根据首尾端点确认的骨架直线,确定与其平行并与烟叶边缘相切的两条直线,两条切线间的距离为烟叶的宽度。实验结果为烟叶宽度测量误差率低于1%,各角度测量数据方差最高为0.0649,表明该算法具有旋转矢量不变性的特点,能够准确提取烟叶的宽度。
According to the problem of low accuracy and inadaptation to the random angle in extraction algorithm of tobacco leaf,a new width extraction algorithm based on rotation vector invariance is proposed.The algorithm uses the minimum exterior rectangle method to obtain the two ends of tobacco leaf at first,then two straight line that paralleling to the skeleton line confirmed by the two end,and tangent to the edge was confirmed,the distance between two tangent line was the leaf width.The experiment result shows that the error rate of tobacco leaf width is lower than 1%,the maximum-variance of measurement angle data is 0.0649,indicates that this algorithm has the characteristics of rotation vector invariance,can accurately extract the width feature of tobacco leaf.