【作者】 谭昌伟；

【导师】 王纪华； 郭文善；

【作者基本信息】 扬州大学 ， 农业信息化技术， 2010， 博士

【摘要】 小麦、水稻是我国、也是江苏省最主要的两大粮食作物,稻麦调优栽培技术在实际生产中发挥了重要作用,但稻麦生产状况的大面积监测预报技术至今较为滞后,遥感技术可以瞬时大面积地同步对稻麦的生长状态和生长环境进行监测预报,为实现大面积、低成本早期监测预报稻麦品质和产量提供了技术保障。针对当前大田稻麦调优栽培技术单一、栽培模式繁多等这些亟待进一步探讨和需要解决的实际问题,同时为快速、无破坏地掌握区域性稻麦长势信息,满足稻麦实际生产管理需求,最终实现优质、高产、高效、安全、生态这一稻麦种植生产目标,本研究围绕稻麦长势参数、小麦产量及主要籽粒品质参数与遥感变量间定量关系,以遥感信息为技术支撑,研究遥感技术监测稻麦主要调优栽培指标的机理和方法,探讨应用遥感技术监测稻麦长势、预测小麦籽粒品质和估算小麦产量的可行性。主要研究内容和结果如下:(1)基于Landsat TM数据的小麦长势状况遥感监测研究分析不同生育时期小麦理化参数间的相关性,结果表明,叶片氮含量与籽粒蛋白质、湿面筋及淀粉间均呈显著或极显著相关,但以开花期关系最为密切,说明开花期可作为遥感预测籽粒品质更为理想的时期。分析小麦遥感变量与长势参数间的关系,结果表明,遥感监测小麦SPAD、生物量、LAI、叶片氮含量和叶片含水量时,拔节期,分别选用B5、NDVI、DSW5、B2和RVI作为敏感遥感变量,开花期,分别选用NRI、B4、NDVI和NDWI2作为敏感遥感变量,经评价后,确定了拔节期和开花期小麦长势遥感监测模型。利用上述模型,输入由目标影像生成的遥感光谱变量图并进行解算,再叠加小麦种植分布图和行政边界矢量数据,参照长势参数等级标准,生成LAI、SPAD、生物量和叶片氮含量遥感监测专题图,并以LAI为主,生成拔节期和开花期具有实际农学意义的区域性小麦长势分级图。(2)基于Landsat TM数据的小麦籽粒品质和产量遥感监测预测研究分析小麦叶片氮含量与主要籽粒品质指标间、籽粒蛋白质含量及产量与遥感变量间的相关性,结果表明,开花期利用NDVI预测籽粒粗蛋白质含量和产量是最合适的。采用间接模式,利用遥感监测开花期叶片氮含量,结合叶片氮含量与籽粒粗蛋白质含量间的定量关系,构建及验证直接模式和间接模式下籽粒粗蛋白质含量和产量遥感预报模型,且精度较高。依据籽粒蛋白质含量和产量等级划分标准,生成不同等级籽粒蛋白质含量和产量卫星遥感预报图,以及江苏省小麦品质区划图,为相关部门及时准确地提供小麦籽粒品质和产量信息,利于企业节本增效。(3)基于地面光谱数据水稻理化参数的遥感监测研究分析水稻叶片光谱特性,结果表明,水稻叶片反射光谱与透射光谱的波形特征及趋势极为相近,而与吸收光谱相反,可见光波段因受到色素吸收的影响,导致光谱反射率和透射率都较低;近红外平台波段因受到叶片内部的多次光散射影响,导致光谱反射率和透射率都趋于50%,而吸收率极低。分析遥感光谱变量与水稻理化参数间的相关性,结果表明,在400-1250nm波段,可以利用叶片光谱反射率诊断LAI、SPAD和CHL,在1350-1550nm波段,诊断LWC和LNC;利用NDVI、NDWI、GreenNDVI和NRI等光谱植被指数诊断相应的叶片营养及LAI是可行的,并构建及验证了以这些遥感参量为自变量的遥感诊断模型,尤其利用NDWI(860, 1240)诊断LWC、GreenNDVI诊断SPAD更为理想,构建的诊断模型分别为:y = 74.39x + 71.702和y = 64.737x + 25.221。(4)水稻氮素营养高光谱遥感监测研究通过分析遥感光谱变量与水稻氮含量间的关系,结果表明,以796nm附近处的光谱反射率为自变量的氮含量线性模型用于诊断水稻氮素营养是可行的,以738nm附近处的一阶微分光谱反射率为自变量的氮含量指数模型诊断水稻氮素营养亦是可行的,且优于以796nm附近处的光谱反射率为基础构建的线性模型,经预测性评价后,最终提出了采用以植被指数的归一化变量(SDr-SDb)/(SDr+SDb)为自变量构建的水稻氮素营养高光谱遥感诊断模型(y = 365.871 + 639.323(SDr-SDb)/(SDr+SDb))诊断水稻氮素营养是最理想的。(5)调优栽培集成技术研究分析小麦拔节期叶片SPAD值与施氮量间的定量关系,根据小麦叶片SPAD值的遥感监测结果,有针对性地提出某一区域的氮肥施用量,再根据某一区域内不同地块的小麦叶片SPAD值,提出单个地块的氮肥施用量,综合上述分析,已初步实现卫星遥感与小麦调优栽培技术的集成,形成了一套基于卫星遥感和叶片SPAD值的小麦实时诊断和变量施肥技术,这项技术可用于小麦长势实时诊断和变量施肥调控。将卫星遥感影像与水稻调优栽培技术进行集成,依据LAI等级划分标准,生成水稻长穗期长势状况遥感监测分级图,并依据对水稻长穗期长势状况卫星遥感监测结果,集成水稻调优栽培技术,提出水稻长穗期生长异常的具体调控措施,建立一套基于卫星遥感与水稻长穗期LAI的调优栽培技术,这项技术可用于针对遥感监测长势状况提出配套的栽培管理措施。更多还原

【Abstract】 Wheat and rice are the two most important food crops in our country and Jiangsu province. The optimized cultivation techniques have played an important role in the actual production of wheat and rice. However, the technology of monitoring and forecasting wheat and rice production status on large areas is the more lag so far. The growth status and environment of wheat and rice can be monitored and forecasted synchronizely at a large-scale and in real-time by remote sensing technology. It achieved to provide the technical support for large-scale, low-cost monitoring and forecasting of early quality and yield of wheat and rice.The study is in connection with the practical problems that the currently rice-wheat cultivation techniques imperfect which needed further development and solved, at the same time,it need grasping information on the regional rice-wheat crop fastly and without damage and meeting the demand of rice and wheat production management practice, and ultimately making the target of high-quality, high yield, efficient, safe, ecological cultivation of this rice-wheat production come out, Around the variables relationship between the growth of rice-wheat crop, wheat yield and the main grain quality parameters and remote sensing quantitative, this study depends on the remote sensing technology to study of tuning the mechanism and methods of integration between remote sensing technology and rice-wheat cultivating management measures. We also explore the application of remote sensing techniques to monitor the growth of rice-wheat crop, prediction of wheat grain quality and the feasibility estimating of wheat production. The main research content covers the following aspects:(1) The study on monitoring the growth situation of winter wheat by remote sensing based on Landsat TM dataWe analysisd the correlation between physical and chemical parameters different growth stages of wheat.The results showed that leaf nitrogen content and grain protein, wet gluten and starch significantly or very significantly correlated. But the correlation of flowering is most closely.It shows that flowering can be used as the desirable period of time for grain quality forecasted by remote sensing.By analysising of remote sensing variables and are the growth of wheat, the relationship between parameters showed that when monitoring wheat SPAD, biomass, LAI, leaf nitrogen content and leaf water content,in the jointing stage, we selected to B5, NDVI, DSW5, B2 and the RVI as sensitive remote-sensing variables.In the flowering period, we selected NRI, B4, NDVI, and NDWI2 as sensitive remote-sensing variables. After evaluation, the remote sensing monitoring model about the jointing stage of wheat crop is growing and flowering stages has been determined.With the use of the above model, we enter by the spectrum of remote sensing to map that is generated target image and solver variables, and then overlay maps of wheat and administrative boundary vector data, by taking into account parameters of grading standards are growing, generate the matic remote sensing monitoring of map of LAI, SPAD, biomass and leaf nitrogen content.It generated graded map of jointing stage and the flowering period of practical significance of the regional agricultural crop mainly with LAI.(2) The study on monitoring and forecasting wihter wheat grain quality and yield by Remote Sensing based on Landsat TM dataBy Analysising of leaf nitrogen content of wheat grain quality indicators of the main, the grain protein content, yield and remote sensing variables, the correlation between variables, results showed that in the flowering period by the using of NDVI predicted crude protein content and yield of grain is the most appropriate.With the Indirect model, the use of remote sensing to monitor flowering leaf nitrogen content,combined with leaf nitrogen content and grain crude protein content of the quantitative relationship .We constructed and validated the prediction model of remote sensing the direct mode and indirect mode about grain crude protein content and yield, which was high precision. Bsed on grain protein content and yield grading standards, forcasting maps and the zoning map of wheat quality in Jiangsu Province has been generated about different levels of grain protein content and yield,in order to provide wheat grain quality and yield informaition to relevant departments to help enterprises to an abriged version efficiency.(3) Diagnosing leaf nutrition concentration and leaf area index in rice through leaf spectra dataThe research analysised spectral characteristics of rice leaves showed that the rice leaf reflectance spectra and transmission spectra of the waveform characteristics and trends were very similar, while contrary with absorption spectra, due to the impact of the visible light absorpted by the pigment.It resulted in spectral reflectance and transmittance are comparatively low ; near-infrared bands due to being blade platform, the internal effects of multiple light scattering, resulting in spectral reflectance and transmittance have tended to be 50%, while the low absorption rate.Through analysising the correlaition between remote sensing spectral variable physical and chemical parameters in rice. The results showed thatwe can make use of leaf spectral reflectance diagnosis LAI, SPAD and the CHL in the 400-1250nm band, diagnosis, LWC, and LNC in the 1350-1550nm band; It is feasible to use NDVI, NDWI, GreenNDVI and the NRI and other spectral vegetation index in diagnosis of the corresponding leaf nutrition and LAI.We have constructed and validated model of remote sensing variables that using these remote sensing parameters for self-diagnosis.In particular, the use of NDWI (860, 1240) diagnosis of LWC, GreenNDVI diagnosis of SPAD is desirable to construct the diagnostic model, respectively: y = 74.39x + 71.702 and y = 64.737x + 25.221.(4) Hyperspectral remote sensing diagnosis models for nitrogen nutrition status in rice The purpose of study was to analysis the correlation of remote sensing spectral variables and nitrogen content of rice .The results showed that it is feasible. We use the spectral reflectance near to 796nm as independent variables of nitrogen content in the linear model to diagnose rice nitrogen nutrition. And then, near to 738nm Department’s first-order differential spectral reflectance as the independent variable nitrogen content of rice nitrogen nutrition index of model-based diagnosis is also feasible and superior to the spectral reflectance the linear model built which based on at 796.7nm. After forecasting evaluation , the final proposed to adopt a vegetation index, normalized variables (SDr-SDb) / (SDr + SDb) as independent variables built Hyperspectral remote sensing of rice nitrogen nutrition diagnosis model (y = 365.871 + 639.323 (SDr-SDb) / ( SDr + SDb)) diagnosis of nitrogen nutrition of rice is the best.(5) The study on integrating the optimized cultivation techniquesBy the analysis of jointing stage of wheat leaf SPAD values and the quantitative relationship between the amount of nitrogen, we have found the amount of nitrogen fertilizer in a targeted manner of a certain region according to wheat leaf SPAD value of the remote sensing monitoring results. Then according to the different plots within a region Wheat leaf SPAD value, we put forward a single block of nitrogen fertilizer. The above analysis has been initially achieved tuning of satellite remote sensing and wheat cultivation techniques of integration.Then we formed a set of satellite-based remote sensing and leaf SPAD value of the real-time diagnosis and variable fertilization of wheat technology, which can be used for real-time diagnosis of wheat crop is growing and the regulation of variable rate fertilization.Through integrating the satellite remote sensing images and rice cultivation techniques, we generated a long-heading stage of rice crop growth status of remote sensing classification map that based on LAI grading standards.And based on a long-heading stage of rice crop growth status of satellite remote sensing monitoring results, a rice tune excellent cultivation techniques was integrated.Long-heading stage of rice abnormal growth of the specific control measures have been figured out. The establishment of rice tuning cultivation techniques which was satellite remote sensing -based and LAI, the technology can be used for remote monitoring of bloom conditions proposed by supporting cultivation management measures.更多还原