基于正交试验和神经网络的烤烟栽培技术优化

Optimization of Flue-cured Tobacco Cultivation Technique Based on Orthogonal Test and Neural Network

应用正交设计方法,研究烤烟主要栽培因子有机肥比例、种植密度、留叶数与烟叶品质的关系,利用BP神经网络建立了烤烟在豫中地区土壤中等肥力条件下的高产优质栽培模型,最后得出以优质烟内在质量为目标函数的最佳栽培方案,即:有机肥比例25%,种植密度16500株/hm2,留叶数为22片/株。

This article used the orthogonal design method to study the relationship between tobacco quality and main factors of flue- cured tobacco cultivation （the proportion of organic fertilizer, planting density, remained leaves）, and used BP neural network to construct flue - cured tobacco high - yielding and high - quality cultivation model on the middle fertile soil in the central area of Henan province. Taking the internal quality of high - quality tobacco as target function, the best cultivation scheme was finally obtained as follows： the proportion of organic fertilizer being 25%, planting density being 16500 plants/hm^2, remained leaves being 22 leaves per plant.