水保所知识产出(1956---)知识库

针对目前植物叶片几何参数计算中的一些不足，提出适用于离体小麦叶片叶面积、叶长、叶宽和周长的计算
方法。首先使用直方图阈值法对特定背景下小麦图像进行分割；然后采用统计像素数法计算小麦叶面积，采用叶
片边缘相邻像素间距离的方法计算周长，使用改进后的最小外接矩形法计算叶长和叶宽；最后对计算结果与实测
值进行比较，并使用统计分析方法对计算结果进行检验。结果显示，计算值接近实测值，叶面积、叶长和叶宽的均
方根误差（root mean square error, RMSE）分别为0.49 cm 2 、0.57 cm、0.04 cm，决定系数（R 2 ）分别为0.99、0.97、0.96。
RMSE值接近于0、R 2 值接近于1，以及计算值与测量值之间存在的较小差异，表明改进后的叶长和叶宽算法的计算
结果不受小麦叶片形状的影响，方法普适性好，且叶长、叶宽、叶面积和周长算法对相应参数的计算结果准确度高。

Aiming at some deficiency of geometrical parameter calculation in plant leaves at present, we
proposed a computational method suitable for measuring area, length, width and perimeter of excised wheat
leaves. Wheat image was segmented under a certain background by histogram algorithm. Leaf area was calculated
by statistical image number method, and leaf perimeter was calculated on the distance between adjacent pixels of
the blade edge. Leaf length and width were measured by improved minimum enclosing rectangle. Then the
calculated results were compared with the measured values, and statistical analysis was carried out. The results
showed the calculated results were close to the measured values. The root mean square error (RMSE) values for
area, length and width of wheat leaves were 0.49 cm 2 , 0.57 cm and 0.04 cm, respectively. The coefficient of
determination (R 2 ) for area, length and width of wheat leaves were 0.99, 0.97 and 0.96, respectively. The
differences in RMSE values were close to 0 and R 2 values were close to 1, and the calculated results and the
measured values were insignificant, indicating that the improved algorithms for measuring area, length, width and
perimeter of wheat leaves are not influenced by wheat leaf shapes, suitable for different conditions, reliable and
practicable.