Abstract:
Objective Mathematic models for estimating the morphological characteristics of Pennisetum giganteum leaves were constructed to facilitate the utilization of the plant material.
Method Measurements on the length, width, area, and fresh weight of P. giganteum leaves were collected. Specifically, 7 parameters including length, width, (length× width), (length – width), (length + width), (length)2, and (width)2 of the leaves were used to construct fresh weight estimation models using univariate linear regression, binary linear regression, and quadratic trinomial analyses.
Result The correlations among the various measurements had coefficients ranging from 0.74 to 0.91. The estimation models with the highest R2 = 0.99 was on yA = 0.72xP, where yA = leaf area and xP = (length×width), and R2=0.96 on yF=0.02xP, where yF = fresh weight and xP = (length×width). The verification test on the regression equations between the estimated and actual leaf area showed R2 = 0.99 with an RMSE of 18.67 cm2 for leaf area and an RMSE of 0.67 g for fresh leaf weigh.
Conclusion The models for estimating the area and fresh weight of P. giganteum leaves had high degree of validity in providing basic information for Juncao research.
下载: