Abstract:

Objective: Phyllostachys edulis leaves are small and easy to be curled, which increases the difficulty and error of area measurement. Our aim of this study is to build a quick, convenient and accurate measurement of leaf area for P. edulis. Method: We collected leaves from 7 provinces and classified the leaf shapes. The data of leaf length and leaf width, as well as leaf area measure by the scanner were used to build the leaf models. The root mean square error, chi-square test, chi information criterion and predicted precision were used to test the errors and the goodness of the models. Result: 1) According to the length-width ratio, P. edulis leaves can be divided into three categories, which are triangular leaves (length-width ratio ≤ 7.2), elliptic leaves (7.2 < length-width ratio ≤ 8.3) and slender leaves (length-width ratio ≥ 8.3), respectively; 2) Morphological correlation of P. edulis showed that the product of leaf width and length was the closest factor related with leaf area, with ahigh correlation coefficient of 0.993; 3) The Classified fitting model of P. edulis leaf shape was based on the product of leaf length and width, with the determination coefficient of 0.990 1, the root-mean-square error value of 0.159 6, the chi-square value of 10.368 1, the Chi information criterion value of -6 317.10, and the prediction accuracy of 97.73%, respectively. The prediction results were better than that of area measured bya leaf area meter and predicted by whole fitting. Conclusion: The leaf area fitting model based on leaf morphology classification can accurately predict the leaf area by measuring the length and width of P. edulis leaves. In this way, it can not only measure the leaf area more accurate than leaf area meter and the whole fitting method, but also measures fast and conveniently. This model solves the difficulty in measuring the leaf area of P. edulis.

Key words: classified fitting, leaf area meter, scanning, leaf shape, length-width ratio