欢迎访问《麦类作物学报》编辑部网站

作者	单位
赵春奇，李振海，杨贵军，段丹丹，赵钰，杨武德	（1.山西农业大学农学院，山西太谷 030801；2.国家农业信息化工程技术研究中心，北京 100097；3.农业部农业信息技术重点实验室，北京 100097；4.北京市农业物联网工程技术研究中心，北京 100097）

摘要点击次数: 791

全文下载次数: 603

中文摘要:

为优化冬小麦籽粒蛋白含量（GPC）的遥感预测模型，基于2012-2013、2014-2015和2017-2018年冬小麦生长季的田间试验，以植株氮代谢过程及GPC形成规律为依据，构建“植被指数（VI）-农学参数-GPC”的半机理模型，并在此基础上通过引入筋型修正系数λ优化“PNC-GPC”模型，修正小麦筋型对模型的影响，进一步提高“VI-PNC-GPC”模型的精度。结果表明，选取的VI与植株氮浓度（PNC）均极显著相关，其中比值光谱植被指数（RSI）与PNC的相关性最高，相关系数达到0.777，建立的PNC估算模型的决定系数（r）达到0.604，验证nRMSE为9.93%；构建的PNC-GPC模型为GPC=(5.843×PNC+4.847)×λ，r=0.792，验证nRMSE为7.43%；对比不考虑冬小麦筋型的“RSI-PNC-GPC”模型，其r提高了0.145，验证的nRMSE降低了0.86%。综合来看，以PNC为中间变量，通过考虑不同筋型的差异构建的筋型修正系数可以更加准确地预测GPC。

英文摘要:

Remote-sensing estimation of grain protein content(GPC)in winter wheat provided an important basis for winter wheat production and quality classification.Based on field experiments of winter wheat during the growing seasons of 2012-2013, 2014-2015 and 2017-2018, a semi-mechanism model of “vegetation index(VI)-agronomic parameter-GPC” was constructed based on the nitrogen metabolism and GPC formation and the“PNC-GPC” model was optimized by introducing an optimized coefficient(λ) to correct the influence of different gluten types.The ratio spectra index(RSI) had highest correlation coefficient with PNC(r=0.777). Based on RSI, the rof the PNC model reached to 0.604 in the calibration model, and the nRMSE was 9.93% in validation. The “PNC-GPC” model was constructed as GPC=(5.843×PNC+4.847)×λ, r=0.792, and the nRMSE was verified to be 7.43%.Compared to the “RSI-PNC-GPC” model without considering the gluten type of winter wheat, the rof GPC considering gluten type in the study was increased by 0.145 and the verified nRMSE was reduced by 0.86%.Therefore, it is feasible to estimate the GPC with PNC as the intermediate variable. By considering the gluten type of winter wheat, GPC can be predicted more accurately.

查看全文查看/发表评论下载PDF阅读器

关闭