对建立遥感估产模式的几点初步认识

本文从分析遥感光谱参数的生物学意义着手，论证了正确建立遥感估产模型的可能途径。对几种有代表性的遥感估产模型作了分析，作者认为把可见光、近红外波段的遥感信息与热红外信息有机结合是解决遥感估产模型的最佳方案。对ＮＯＡＡ－ＡＶＨＲＲ的第１通道与第２通道光谱数值进行非朗伯体特性的纠正是必要的。遥感估产模型不仅可以使估产的空间尺度大大缩小而且参数数目亦可大大减小，更有利于实际运行。     

基于遥感空间特性的广东省水稻单产快速预估

分析了目前水稻遥感估产的技术现状,基于遥感数据的空间特性,提出了一种快速预测水稻单产的方法,估产试验表明该方法简单实用,具有推广意义.     

ASAR数据与水稻作物模型同化制作水稻产量分布图

提出了利用雷达数据进行水稻估产的技术方法,并以ASAR数据为例,探讨了雷达数据在水稻估产中的可行性.首先利用ASAR数据进行水稻制图,从各时相ASAR数据中提取水稻后向散射系数.随后,基于像元尺度,采用同化方法,以LAI为结合点,将水稻作物模型ORYZA2000与半经验水稻后向散射模型结合,建立嵌套模型模拟水稻后向散射系数.选择水稻出苗期和播种密度为参数优化对象,利用全局优化算法SCE-UA对0RYZA2000模型重新初始化,使模拟的水稻后向散射系数值与实测值误差最小,并由优化后的ORYZA2000模型计算每个像元的水稻产量,生成水稻产量分布图.结果表明,水稻产量分布图能够描绘研究区水稻实际产量的分布趋势,但由于采用潜在生长条件模拟,模拟的水稻平均产量比实测平均值高约13%,验证点的水稻产量模拟值与实测值相对误差为11.2%.由于半经验水稻后向散射模型存在对LAI变化不够敏感和对水层的简化处理,增加了水稻估产的误差.但从总体上看,利用该方法进行区域水稻估产是可行的,并为多云多雨地区的水稻遥感监测提供了重要参考.     

水稻遥感动力估产模拟初探

将遥感手段和动力估产模型相结合，即通过ＮＯＡＡ卫星获得有效绿度模式；通过绿度－叶面积关系式，估算水稻群体叶面积指数；依据水稻生物量分配规律及环境条件对其影响，估算水稻各器官的干重，取得了较好的模拟效果。     

大面积小麦遥感估产模型的构建与调试方法的研究

本文重点研究大面积冬小麦遥感估产模型构建及其调试方法。通过分析冬小麦生长发育过程,对光、温、水、肥等必须条件需求规律研究的基础上,提出了以绿度指数、温度和绿度变化速率等因子,构建大面积冬小麦遥感估产模型。为了适应大面积遥感估产运行系统的需要,在变量获取及模型调试等方面进行了一些新的探索。     

江苏省水稻长势遥感监测与估产

根据1990年以来江苏省进行的水稻长势遥感监测与估产的基础理论、技术方法、长势监测与估产的结果,提出建立省级主要农作物遥感监测运行系统的条件已经具备。它很可能像气象卫星天气预报那样成为第二家为政府和百姓日常工作与生活服务的遥感系统。     

基于多时相MODIS-RVI的玉米遥感估产研究

为了获得更加宏观高效的农作物估产模型,以吉林省德惠市为研究区,以MODIS为数据源,进行了玉米估产模型研究。通过分析比值植被指数(RVI)与玉米产量之间的相关关系,建立玉米单产预测模型。研究表明,利用多时相的RVI对玉米点进行遥感回归估产可得到较好的估算效果,模型相关系数可达0.825,均方根误差为7.61,验证点的实际产量与理论产量间的相对误差均在10%以内,对吉林省德惠市玉米估产模型研究具有一定的指导意义。     

稻田光谱与水稻长势及产量结构要素关系的研究

本文通过对水稻不同生育期的稻田光谱与水稻长势和产量结构的相关分析得出以下几点结论:(1)稻田光谱与水稻长势的相关性较好,特别是在水稻生长后期;(2)抽穗期,稻田光谱与决定水稻产量的各结构要素之间相关系数较高;(3)在水稻灌浆期,稻田光谱与水稻理论产量的相关性较好,特别是800nm的反射光谱值:(4)水稻估产模型是:Y=6.65PVI_D+8.19PVI_H+4.48PVI_Mt+4.36PVI_Ms-0.3,其中:Y-水稻产量;PVI-垂直植被指数,D,H,Mt,Ms,分别代表分化期、抽搐期、灌浆期及乳熟期。     

基于RS和GIS的农作物估产方法研究进展

实时获取农作物长势及产量等信息对于现代农业的发展具有重要意义。近年来,随着遥感技术(remote sensing,RS)和地理信息系统(geographic information system,GIS)广泛应用于农作物估产领域,相继出现了一些较为实用的估产方法,主要有结合辅助数据的估产方法、基于植被指数的估产方法、基于特定模型的估产方法和基于农作物估产平台(软件)的开发等。其中,基于植被指数的估产方法又分为单一和多植被指数估产2类方法。在对近年来该领域大量文献深入研究的基础上,着重就几类热点方法展开论述,并对每类方法的优势和缺陷进行了评述,最后对该领域需要进一步研究的方向进行了探讨和展望,以期为后续研究提供参考。     

二轴土壤背景纠正的植被指数及其在华南水稻遥感估产中的应用

目前国内外学者提出了各种植被指数来进行作物遥感估产的定量研究。这些指数多是基于“土壤线”的存在来进行土壤背景消除的。但它们只消除了土壤背景中的含水量(沿“土壤线”方向)对遥感数据的影响,而没有消除由于不同土壤质地的变化(垂直于“土壤线”方向,如红壤、棕壤等不同的土壤类型)所造成的遥感数据的偏移。本文首次提出了能基本上完全消除土壤背景影响(包括土壤含水量、土壤类型等)的二轴土壤背景纠正的植被指数(TWVI)模型。该指数比目前使用的其它植被指数更适合于作为进行全球监测的植被指数。已成功地应用于华南地区的水稻遥感估产试验。     

水稻微波后向散射系数的模拟分析

提出了一套完整的水稻一次后向散射作用物理模型,通过该模型可以定量地模拟水稻对入射电磁波的后向散射作用,包括不同入射角、不同时相、不同波段、不同极化等各种情况,从而得到大量有价值的模拟结果。通过深入分析这些结果,可以对如何利用ＳＡＲ遥感数据更准确、更经济、更方便地进行水稻识别、长势监测及产品评估等工作提供理论依据和方法指导。     

Automatic Detection of Rice Disease Using Near Infrared Spectra Technologies

The rice disease is one of the most serious injurious factors that cause major loss of rice production and subsequent economy in agricultural industry. This study explored a new method for obtaining information of the rice disease in a short term through model regression methods. The spectrum characteristics of rice leaves under different disease damage were firstly analyzed for its relationship with rice disease level. The sensitive bands of the spectrum for accurately supervising rice diseases were selected with principal component analysis (PCA). The stepwise regression method and BP neural network were both used to establish the spectrum-based models for recognizing rice diseases. Results showed that five major characteristic bands were determined by PCA (990, 1850, 660, 1921, and 1933 nm) for monitoring foliar rice diseases, among which the edge area for red light had the best correlation with rice disease level was also selected as the parameter to establish the model. Specifically, the composite reflectivity of wavelengths between 990 and 1933 nm was negatively related to rice brown spot diseases stress, which was then used to establish the model. Parameters of the red edge area and the ranged reflectivity between 660 and 990 nm were used to establish models for monitoring rice sheath blight diseases. Totally, there were 60 samples employed to build models for identifying the two diseases by the stepwise regression method and the BP neural network method, and the rest 41 ones were used for further model verification. Compared with the stepwise regression analysis, BP neural network was evaluated to perform better with characteristic bands at 660, 990, and 1933 nm. In conclusion, the establishment of the function model in our study can be implemented to monitor rice diseases, which provided a theoretical basis for indirect and rapid monitoring rice diseases.     

SO2污染环境下水稻导数光谱与生理生化指标的相关性研究

通过田间开顶式小区熏气试验,研究在SO2急性伤害条件下水稻冠层导数光谱与叶片含硫量、叶液pH值以及叶绿素含量的相 关性。分别选择分蘖期和抽穗期显著相关的波段(分蘖期： 689 nm、584 nm、570 nm; 抽穗期： 689 nm、584 nm、585 nm)建立 预测叶片含硫量、叶液pH值及叶绿素含量的回归模型,并分别用拔节期和灌浆期相应导数光谱反射率检验模型预测精度。结果表明 ,由分蘖期建立的回归模型估测拔节期叶液pH值以及叶绿素含量与实测值之间相关系数分别为0.884和0.630; 由抽穗期建立的回 归模型估测灌浆期的叶片含硫量、叶绿素含量与实测值之间相关系数分别为0.659和0.768,均通过显著检验。     

Identification of Aman Rice Crop in Bangladesh Using Temporal C-Band SAR – A Feasibility Study

In-season rice area estimation using C-band Synthetic Aperture Radar (SAR) data from RADARSAT-1 is being done in India for more than a decade. Decision rule based models in backscatter domain have been calibrated and validated using extensive field data and a long term backscatter signature bank of rice fields has been developed. Since the rice crop growing environment in India is a diverse one in the world having all the rice cultural types, the rice backscatter is quite exhaustive. This paper highlights the results of classification of rice lands in Bangladesh using the signature bank of India. The results showed that the Aman rice crop of Bangladesh has a typical temporal backscatter of shallow and intermediate rice fields of that of West Bengal state. The mean backscatter of the intermediate/deep water fields in southern Bangladesh was ?19?dB, while that of shallow cultural types mostly in northern Bangladesh was ?17?dB. The signature of the rice crop in Southern Bangladesh matched well with that of Gangetic West Bengal, particularly that of the 24 Parganas, Howrah and Hughli districts. The signature of rice crop in the Sub-Himalayan West Bengal particularly that of Dinajpur and Maldah districts matched well with that of the northern area of Bangladesh. State level rice area estimated using the selected models was found with in 5% deviation from that of the reported acreage.     

District wise yield models of rice in Bihar based on water requirement and meteorological data

Pre-harvest crop production forecast has been successfully provided by remote sensing technique. However, the probability to get cloud-free optical remote sensing data during kharif season is poor. Microwave data having the capability to penetrate cloud is used in the absence of cloud free optical remote sensing data. Yield models in broad band frequency range are in development stage. Meteorological yield models are developed and predicted yield is combined with area estimated by remote sensing data to provide rice production forecast. This paper describes the methodology adopted for improving the predictability of rice yield before harvest of the crop in Bihar province by taking into consideration meteorological parameters during its growth cycle upto October. Models developed using fortnightly meteorological data have been found to give reasonably fair indications of expected yield of rice in advance of harvest. The yield predictions have been made based on meteorological data and effective rainfall based on water requirement calculations representing a group of districts under similar agro-climatic zones, which could be further improved by incorporating meteorological data of individual districts within each group.     

水稻叶面积指数的高光谱遥感估算模型

通过不同氮素营养水平的水稻田间试验 ,采用单变量线性与非线性拟合模型和逐步回归分析 ,用1 999年试验数据为训练样本 ,建立水稻LAI的高光谱遥感估算模型 ,用 2 0 0 0年试验数据作为测试样本数据 ,对其精度进行评价和验证。结果表明 ,高光谱变量与LAI之间的拟合分析中 ,蓝边内一阶微分的总和与红边内一阶微分的总和的比值和归一化差植被指数是最佳的变量     

水稻生长期微波介电特性研究

利用植被介电常数的Debye-Cole双频色散模型，模拟计算了广东肇庆水稻试验区1996年晚稻和1997年早稻人插秧期、发蘖期、扬花期到成熟期各生长期的介电常数值，并根据计算结果，探讨了电磁波频率、水稻含水量、温度、含盐度及水稻冠层干体密度对介电常数的影响。其中，不同生长期水稻的介电常数各不相同，不同水稻类型（早稻和晚稻），介电常数的变化趋势不尽相同。电磁波频率、水稻含水量、温度和水稻冠层干体密度均对介电常数有不同程度的影响，而含盐度却对介电常数影响不大。     

Multiple forecasts of kharif rice in orissa state-four year experience of fasal pilot study

Considering the requirement of multiple pre-harvest crop forecasts, the concept of Forecasting Agricultural output using Space, Agrometeorology and Land based observations (FASAL) has been formulated. Development of procedure and demonstration of this technique for four in-season forecasts for kharif rice has been carried out as a pilot study in Orissa State since 1998. As the availability of cloud-free optical remote sensing data during kharif season is very poor for Orissa state, multi-date RADARSAT SCANSAR data were used for acreage estimation of kharif rice. Meteorological models have been developed for early assessment of acreage and prediction of yield at mid and late crop growth season. Four in-season forecasts were made during four kharif seasons (1998-2001); the first forecast of zone level rice acreage at the beginning of kharif crop season using meteorological models, second forecast of district level acreage at mid growth season using two-date RADARSAT SCANSAR data and yield using meteorological models, third forecast at late growth season of district level acreage using three-date RADARSAT SCANSAR data and yield using meteorological models and revised forecast incorporating field observations at maturity. The results of multiple forecasts have shown rice acreage estimation and yield prediction with deviation up to 14 and 11 per cent respectively. This study has demonstrated the potential of FASAL concept to provide inseason multiple forecasts using data of remote sensing, meteorology and land based observations.     

引黄灌区水稻叶面积指数的高光谱估测模型

水稻叶面积指数（leaf area index,LAI）是评价其长势的重要农学参数,高光谱遥感能够实现叶面积指数的快速无损监测。为了寻找反演水稻LAI的最优植被指数,扩展水稻LAI高光谱估测模型的普适性,选取宁夏引黄灌区水稻为研究对象,通过设置不同氮素处理,借助相关分析、回归分析等方法研究高光谱植被指数与水稻LAI之间的定量关系,并通过确立的最优波段组合,构建4种植被指数与水稻LAI的高光谱反演模型。结果表明,水稻LAI在抽穗末期达到最大值,并随氮素水平的增加而增加;水稻冠层原始光谱反射率在400~722 nm和1 990~2 090 nm波段与LAI达到极显著负相关水平,在近红外区域760~1 315 nm与LAI呈极显著正相关。模型检验结果表明,以比值植被指数RVI（_850,750）为变量建立的水稻LAI估测模型最佳,研究结果可为水稻LAI的高光谱估测提供地域参考。     

Evaluation of machine learning models for rice dry biomass estimation and mapping using quad-source optical imagery

ABSTRACT

Several machine learning regression models have been advanced for the estimation of crop biophysical parameters with optical satellite imagery. However, literature on the comparative performances of such models is still limited in range and scope, especially under multiple data sources, despite the potential of multi-source imagery to improving crop monitoring in cloudy areas. To fill in this knowledge gap, this study explored the synergistic use of Landsat-8, Sentinel-2A, China’s environment and disaster monitoring and forecasting satellites (HJ-1 A and B) and Gaofen-1 (GF-1) data to evaluate four machine learning regression models that include Random Forest (RF), Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), and Gradient Boosting Decision Tree (GBDT), for rice dry biomass estimation and mapping. Taking a major rice cultivation area in southeast China as case study during the 2016 and 2017 growing seasons, a cross-calibrated time series of the Enhanced Vegetation Index (EVI) was obtained from the quad-source optical imagery and on which the aforementioned models were applied, respectively. Results indicate that in the before rice heading scenario, the most accurate dry biomass estimates were obtained by the GBDT model (R² of 0.82 and RMSE of 191.8 g/m²) followed by the RF model (R² of 0.79 and RMSE of 197.8 g/m²). After heading, the k-NN model performed best (R² of 0.43 and RMSE of 452.1 g/m²) followed by the RF model (R² of 0.42 and RMSE of 464.7 g/m²). Whist the k-NN model performed least in the before heading scenario, SVM performed least in the after heading scenario. These findings may suggest that machine learning regression models based on an ensemble of decision trees (RF and GBDT) are more suitable for the estimation of rice dry biomass, at least with optical satellite imagery. Studies that would extend the evaluation of these machine learning models, to other parameters like leaf area index, and to microwave imagery, are hereby recommended.