20个品种枣树叶面积回归方程的建立

摘要

以延川县20个枣树品种的成熟叶片为材料,采用扫描叶片并结合Image-Pro Plus 6.0程序的数字图像处理技术测定枣叶的长、宽及面积.运用SPSS 17.0统计软件对测量数据进行回归分析,建立了枣树估算叶面积的一元回归方程、二元回归方程、三元回归方程和总的三元回归方程.结果表明,建立的各品种枣树叶面积估算的三元回归方程,是估算其叶面积的最佳回归方程,估算的平均误差率仅为2.06％(1.79％～2.5％)；建立的总的枣树叶面积估算的三元回归方程(y6=-0.023x1-0.335x2+0.674 x3+1.145),可以作为估算所有枣树叶面积的优化回归方程,对20个品种枣树叶面积估测的平均误差率只有3.03％(2.5％ ～3.53％).该研究结果为枣树叶面积的快速测定提供了一个科学有效的方法,在枣树科学研究和生产实践中有较好的应用价值.%The jujube leaves of 20 species from Yanehuan, Shanxi were used as material. The index like length, width and area of leaf was measured by using scanner and combining with Image-Pro Plus 6.0 program of digital Image processing technology. After using SPSS17.0 statistical software to make regression analysis of measured data, unary regression equations, binary regression equations, ternary regression equations and universell adj ternary regression equations were established of estimated jujube trees' leaves area. The results showed that the established ternary regression equations of leaf area estimate for every jujube tree varieties were the best regression equations of estimate leaves area. The average estimate error rate was only 2.06% (1.79% ～2.5%); the established universell adj regression equations (y6 = -0.023x1 -0.335x2 +0.674x3 +1.145) can be used for any jujube trees leaf area estimated optimization regression equation with an average error rate 3.03% (2.5% ～ 3.53%) to the leaves area of 20 varieties jujube trees. The study results provided a scientific and effective method for rapid determination of jujube tree leaves area.