烟株打顶适合度指数模型的构建及应用

Construction and Application of the Suitability Index Model for Topping Tobacco Plants

烟株打顶是烟叶种植过程中的一个重要环节,通过对烟株进行打顶,可人为干预烟株生长过程,调整烟株营养物质的输送位置,促进烟叶的生长发育,从而提高烟叶产量.通过分析烟株打顶时对应烟叶的生长状况数据,对打顶后烟株的生长情况进行统计,探讨影响烟株打顶的关键因素,构建烟株打顶适合度指数模型,并采用BP神经网络对烟株打顶模型进行深入研究.研究结果表明,影响烟株打顶的主要因素包括烟株叶片数、花蕾到烟株最上部叶距离及花蕾开放程度.BP神经网络分析模型验证集准确率为78.33%,R^(2)=0.82,MSE=0.59.采用J2EE平台结合Spring框架对烟株打顶适合度指数模型进行应用开发,经实地验证,模型在实际生产中的平均吻合度为79.32%.研究结果为烟株打顶农事操作提供了合理的判断依据,为烟草种植数字化和管理科学化提供了有力支持.

Topping tobacco plants is an important part of theprocess in cultivation of tobacco.By topping tobacco plants,it is possible to artificially intervene the growth process of tobacco plant,change the delivery destination of nutrients in tobacco plant,and alter the growth and development status of tobacco leaves,thereby increasing the tobacco yield and income for tobacco farmers.This study analyzed the growth status data of tobacco leaves corresponding to tobacco bead topping,statistically analyzed the growth status of tobacco plants after topping,studied the key factors affecting tobacco topping,and constructed a tobacco topping suitability index model.BP neural network was used to study the topping model of tobacco plants.The main factors affecting the topping of tobacco plants have been determined to be the number of leaves,the distance from the flower bud to the topmost leaf of the tobacco plant,and the degree of flower bud opening.The accuracy of the BP neural network analysis model validation set was 78.33%,R^(2)=0.82,MSE=0.59.An application for the tobacco topping suitability index model was developed using the J2EE platform Spring framework.According to on-site verification,the average goodness of fit of modelin actual production was 79.32%.This study provides a reasonable judgment basis for the agricultural operation of tobacco bead topping,and provides scientific support for the digitization and scientific management of tobacco planting.