摘要
本实验采用数码相机拍照,利用CAD图像处理方法得到楸树叶片的面积,并同时测量叶长、叶宽计算出长宽比、长乘宽,通过逐步回归的方法对与叶面积有关的指标进行筛选。用最后得到的最能反映叶面积的指标长乘宽,建立不同叶形的叶片长乘宽与叶面积的一元线形回归预测模型方程,分别是倒心形Y=0.652 0x+16.041,决定系数R2=0.722 1,水滴形Y=0.505 8x+112.47,决定系数R2=0.499 4,三裂Y=0.321 1x+155.98,决定系数R2=0.600 6,分析认为可将叶长乘宽作为重要的叶形指数,进行楸树生长等相关研究。
The experiment took photograph with digital camera, measured leaf area of Manchurian Catalpa by the method of CAD image processing, and in the meanwhile, measuring the length and width of the leaves, then calculated length-width ratio, length multiplied by width, screened the exponents of leaf area by the method of step-by-step regres- sion. Using the numbers of length multiplied by width to set up unitary linearity regression model equation of different leaf shape were respectively inverted heart-shaped Y=0.652 0x＋16.041, coefficient of determination R^0.722 1, drip type Y=0.505 8x＋112.47, coeffieient of determination R2=0.499 4, trilobation type Y=0.321 1x＋155.98, coefficient of determination R2=-0.600 6, analysis suggested that leaf length multiply width was important leaf index to carry out the study on the growth of Manchurian Catalpa.
出处
《甘肃林业科技》
2008年第4期27-29,共3页
Journal of Gansu Forestry Science and Technology
基金
"十一五"国家科技支撑计划项目2006BAD01A1602
2006BAD24B08资助
关键词
楸树
叶面积
CAD图像处理
回归模型
Manchurian Catalpa
leaf area
CAD image processing
regression model
