以HJ-CCD影像为基础的冬小麦孕穗期关键苗情参数遥感定量反演

Quantitative Inversion of Key Seedling Condition Parameters in Winter Wheat at Booting Stage Using Remote Sensing Based on HJ-CCD Images

【目的】应用卫星遥感数据可以及时获取大田种植作物"面状"苗情信息,准确反映作物群体苗情状况及其趋势,服务于产量预报和实际生产。进一步深化冬小麦关键期苗情遥感反演机理与方法,为大田种植管理提供及时信息。【方法】结合2011—2013年定点观测试验,以环境减灾卫星HJ-CCD数据为遥感影像源,着重研究样本实验区孕穗期冬小麦关键苗情参数与籽粒品质参数和产量间及其与卫星遥感变量间的定量关系,进一步增强遥感反演的机理性和重演性,与地面实测结果一起建立模型共同分析,提高遥感反演的定量化水平和可信度;以相关性最高为原则,筛选反演孕穗期冬小麦叶面积指数、生物量、SPAD以及叶片含氮量的敏感卫星遥感变量,并以2013年数据为建模样本、2011年和2012年数据为验证样本,分别构建及评价基于HJ-CCD影像遥感变量孕穗期叶面积指数、生物量、SPAD和叶片含氮量监测模型。【结果】冬小麦处于孕穗期,植被衰减指数(PSRI)可作为反演冬小麦叶面积指数、SPAD和叶片含氮量的敏感遥感变量,比值植被指数(RVI)可作为反演冬小麦生物量的敏感遥感变量,所构建的遥感反演模型是可靠的,且精度较高,尤其利用PSRI反演叶片含氮量最可靠。模型的决定系数(R2)分别为0.651、0.585、0.630和0.675,均方根误差(RMSE)分别为1.344、4.62、0.618%和2 804.3kg·hm-2。以此为依据,为表征该研究的实际农学意义,对冬小麦不同等级的关键苗情参数进行遥感反演并制图分析,从而量化表达了冬小麦关键苗情参数区域空间分布,不仅有助于制定冬小麦田间补救措施和水肥资源调配方案,而且为农业政策的制订和粮食贸易提供决策依据。【结论】构建的冬小麦孕穗期关键苗情参数遥感反演模型是可行的,为大田生产提供了一种快速、便捷、费用低廉的大面积作物苗情参数提取方法,可支持农业研究者、涉农部门领导和种植管理者获取及时有效的农情信息。

【Objective】Application of satellite remote sensing data can timely get field planting crops ‘planar＇ growth information, accurately reflect the situation and trend of crop seedling condition, serve the yield forecast and actual production. The purpose of this research was to deepen the mechanism and methods of remote sensing inversion of winter wheat seedling condition in the key period, and this research will provide timely support information and technology for farm production management. 【Method】Based on experimental data obtained from 2011-2013 in the fixed-point observation experiment, and using HJ-CCD satellite images, the quantitative correlations between key seedling condition parameters of winter wheat at booting stage in sampling regions and the grain quality parameters, production, and remote sensing variables were emphatically analyzed. In order to further enhance the mechanism and reproducibility of remote sensing inversion models, which were built and analyzed with ground measuring results, the quantitative level and reliability of remote sensing inversion models were raised. Models for monitoring the leaf area index, biomass, SPAD value, and leaf nitrogen content of winter wheat at booting stage using remote sensing variables extracted from the HJ-CCD images were built and assessed, respectively. 【Result】It is possible to invert leaf area index, SPAD value and leaf nitrogen content of winter wheat at booting stage by plant senescence reflectance index（PSRI）, and invert biomass by ratio vegetation index（RVI）, respectively. The remote sensing inversion models of the leaf area index, SPAD value, leaf nitrogen content and biomass of winter wheat were credible, and higher precision was obtained with determination coefficient（R2） of 0.651, 0.585, 0.630 and 0.675, respectively, and with root mean square error（RMSE） of 1.344, 4.62, 0.618% and 2 804.3 kg·hm-2, respectively. It was especially reliable to inverse leaf nitrogen content by PSRI. According to the above results, the spatial distribution of the seedling condition parameters of winter wheat could be implemented with agricultural thematic maps of monitoring the key seedling condition parameters at different classes with remote sensing method, thus achieved quantitative expression of regional spatial distribution of the seedling condition parameters. It not only contributes to drawing up the plan of winter wheat field remedial measures and the water and nutrient resources scheduling, but also offers the decision basis for determination of agricultural policy and food trade. 【Conclusion】 The remote sensing inversion models of winter wheat key growth of seedling parameters at booting stage is feasible, and provide a quick, convenient and affordable method to extract the parameters seedling growth of large area for field production. The results of this research can provide timely valuable agricultural information for agronomists, agricultural departments, and farm managers.