CAJ | 학술논문

为了使大豆种植者能够较准确地预测大豆各生育阶段的具体日期,使用BP神经网络构建了简单而有效的预测模型。以播种日期作为模型的主要预测因子,排除病虫害以及突发自然灾害等干扰因素影响,对大豆营养期和生殖生长期分别建模。以垦农18号大豆作为研究对象,以黑龙江九三垦区作为试验基地,历时5年,对大豆的生长阶段数据进行统计。对数据进行归一化处理后,用BP神经网络进行训练和预测,预测的结果表明模型的平均预测精度较高,具有较好的推广应用价值。
위료사대두충식자능구교준학지예측대두각생육계단적구체일기,사용BP신경망락구건료간단이유효적예측모형。이파충일기작위모형적주요예측인자,배제병충해이급돌발자연재해등간우인소영향,대대두영양기화생식생장기분별건모。이은농18호대두작위연구대상,이흑룡강구삼은구작위시험기지,력시5년,대대두적생장계단수거진행통계。대수거진행귀일화처리후,용BP신경망락진행훈련화예측,예측적결과표명모형적평균예측정도교고,구유교호적추엄응용개치。
In order to predict the exact date of soybean growth and development stages for soybean planters,this paper constructed an easy and effective prediction model.Using planting date as main predict factor,it eliminated the effect of plant diseases and insect pests and natural disaster,constructed models for vegetative growth stage and reproductive growth stage of soybean respectively.By used Ken Nong 18 as research subject and Jiu San reclamation area as experimental base,it collected soybean growth data that lasted 5 yesrs.After normalizing data,using BP NN training and testing the data.The result showed the model had good average prediction accuracy and indicated that the BP NN method could be effectively applied in crop modeling.