基于机械收粒的玉米籽粒水分预测模型及其应用

In order to solve the problems of determining the best harvest time of mechanical harvesting maize grain, the multi-point experiments using different types of maize hybrids will be conducted in the three main maize producing regions which are northwest, northeast and Huanghuaihai in China. The characteristics of maize grain dehydration of different types of hybrids after physiological maturity in different ecological environments and the key influencing factors will be studied. The relationship between grain moisture content of different types of maize hybrids and mechanical harvesting qulity index and the most suitable grain moisture content range of different types of hybrids for mechanical harvesting grain will be determined through the harvesting quality evaluation at different grain moisture content levels. The prediction model of grain moisture content will be established based on the quantitative relationships between the key ecological factors, cultivation measures and grain moisture content by using statistics, mathematics and artificial intelligence methods based on the resent 20 years' meteorological factors (1993-2013). The accuracy and applicability of the prediction model will be tested by using empirical method in field experiments. The forecasted and recommended expert system of appropriate maize harvest data for mechanical harvesting grain will be developed based on the resent 20 years' meteorological data (1993-2013) from national meterological stations and by considering the demands of maize purchase and processing. The recommended expert system will be validated and applied to maize fields in the three main ecological regions.

为了解决机收粒玉米最佳收获期确定的问题，通过在西北、东北和黄淮海三大玉米主产区设置多点联网试验，研究不同类型玉米品种在典型生态区玉米生理成熟后籽粒脱水特征及其关键影响因素，通过籽粒不同水分含量下的机械收收获籽粒质量评价，明确不同类型品种籽粒含水量与机械收获质量指标的关系，确定不同类型玉米品种适宜机械收获籽粒的水分分布范围；根据关键生态气象因子、栽培措施与籽粒脱水速率的定量关系，应用统计学、数学和人工智能方法建立玉米后期籽粒含水量预测模型，通过实证方法检验模型的准确性和适用性。利用近20年（1993年-2013年）全国各站点气象数据，综合考虑企业收购价格与烘干成本，开发玉米机收粒适宜收获期预测推荐专家系统，并在各生态类型区进行大田验证与应用。

通过在黄淮海夏玉米区和东北、西北春玉米区开展不同品种、不同栽培措施试验，系统研究了三大玉米产区玉米自籽粒形成直至生理成熟后一段时间内籽粒脱水特征及其影响因素。结果表明，不同玉米产区，自玉米籽粒形成期开始，籽粒含水率接近90%，随着生育进程推进，直到生理成熟期，籽粒含水率呈不断下降的变化趋势，生理成熟后随着在田间站秆时间的延长，籽粒含水率进一步下降并趋于不变；黄淮海夏玉米区现阶段代表性品种，生理成熟期籽粒含水率在21.5-33.1%之间，籽粒含水率降至13-16%后达到平衡水分状态；东北春玉米区，现阶段代表性品种的生理成熟期籽粒含水率在19.3-37.2%之间，当平均气温降至17-18℃时籽粒水分进入缓慢下降阶段，当气温降至7-8℃时籽粒含水率呈现不变趋势，达到平衡水分状态。在西北春玉米区，现有代表性品种生理成熟期籽粒含水率在21.8-37.5%之间。环境因素中，大气温度是籽粒脱水速率的主控因素；植株性状中，苞叶长短、穗轴生理成熟期含水率、果穗夹角、穗粒数、穗长与行粒数比值均显著影响籽粒脱水速率；栽培因素中，品种对籽粒脱水速度的影响最大，种植密度对籽粒脱水速率影响不显著，灌溉量和施肥量及地膜覆盖均影响籽粒脱水速率。研究发现，针对确定品种，从籽粒形成后（含水率为90%时）到生理成熟期，再到成熟后籽粒含水率不变之前，籽粒含水率的变化与吐丝后积温间呈稳定的函数关系，可用Logistic Power模型：Y=90/[1+(x/a)b]预测吐丝后一定积温时的籽粒含水率，且模型中的参数a和b有明确的生物学意义，a为籽粒含水率从最高（90%）降至一半（45%）时所需的吐丝后积温，b为籽粒含水率从45%降至生理成熟期的斜率，反映灌浆后期籽粒的脱水速率。通过大量田间机械粒收数据调查与分析，发现收获期籽粒含水率显著影响破碎率，籽粒17-22%的含水率为最佳收获质量的含水率。基于此，建立了机械粒收最佳收获期推荐专家系统，并在三大玉米产区进行了验证，制作了黄淮海区在小麦种植前籽粒含水率降至28%和25%的区域图，为黄淮海夏玉米机械粒收最佳收获期推荐奠定了基础。