利用多时相NDVI 监测京郊冬小麦种植信息

物候和时相信息在农作物种植信息提取方面有十分重要的应用价值, 利用多时相L andsat TM 数据, 结合冬小麦的波谱和时相信息, 成功提取了北京地区的冬小麦种植信息。首先, 选用了2003 年4 月7 日、5 月1 日、5 月25 日、6 月18 日不同时相的4 景TM 卫星影像, 计算了不同时相的NDV I 时间谱图像数据; 其次, 结合北京地区农作物种植的实际情况, 提取并分析了北京春夏季主要植被地物(冬小麦、苜蓿、苗圃、春玉米、树林等) 的NDV I 时间谱特征; 第三, 利用不同时相的NDV I 图像数据, 通过NDV I 图像通道间的逻辑运算算法, 成功提取了2003 年北京地区的冬小麦种植信息, 提取精度达到96. 92%; 最后, 与2002 年收割小麦的统计数据相对比, 监测了北京各郊区县的冬小麦种植结构调整情况。结果表明, 多时相、多光谱遥感数据在作物种植信息的监测中有十分明显的技术优势和重要的应用潜力。     

基于GF-1遥感数据决策树与混合像元分解模型的冬小麦种植面积早期估算

我国西北地区耕地细碎,冬小麦种植面积提取时混合像元较多,所以将决策树和混合像元分解相结合可大大提高解译精度。以高时间分辨率及较高空间分辨率的GF-1卫星遥感数据为研究数据源。根据冬小麦和其他各类地物在不同时相数据上NDVI值的变化特性及特征值差异,建立决策树模型,快速高效地提取冬小麦像元。运用线性光谱混合模型,降低混合像元的影响,进一步精确提取冬小麦的种植面积。最后与实测样方的冬小麦种植面积数据进行比较,验证提取精度。结果表明:研究区内冬小麦种植面积提取精度达90%以上,Kappa系数接近0.8,可较为准确地反映出区域内冬小麦的分布情况。利用较高分辨率的遥感影像并结合决策树分类和混合像元分解可以较准确地提取耕地破碎地区作物种植面积,对开展早期农作物面积遥感监测有较大帮助。     

HJ星数据在关中冬小麦种植面积遥感监测中的应用 

应用冬小麦抽穗期的环境小卫星影像对关中地区依据行政区划划分,采用多种遥感监测方法进行冬小麦种植面积提取,提取精度达95.64%。结果表明:抽穗期是关中地区提取冬小麦种植面积的最佳时相,同时也表明环境小卫星可以用来监测大范围的冬小麦种植面积,并得到很高的精度。在未来的研究中,对于地形差异较大的地区,可依据地形和冬小麦的种植结构对研究区进行划分,针对每个子研究区采用不同的遥感监测方法以提高提取精度。     

基于遥感影像时间序列的冬小麦种植监测最佳时相选择研究

遥感影像植被分类的最佳时相对作物种植面积遥感监测非常重要。根据2005~2006年北京冬小麦不同物候期的Landsat TM影像和2006年Spot\|2影像,计算了各时期影像中主要植被类型的光谱可分性距离,分析了北京郊区主要植被物候差异和光谱可分性;对各生育期的遥感影像及其主要组合进行了监督分类,采用总体精度和分类效率指标两个参数,结合地面GPS调查数据,对分类结果进行了精度评价。结果表明：北京地区小麦监测最佳时相是4月上旬,影像分类的总体精度为92.9％,明显优于其它单时相影像的分类结果;发现北京郊区冬小麦光谱分类的最佳时相组合为4月上旬(起身期)和5月下旬(灌浆期),分类总体精度为94％。     

基于多时相NDVI及特征波段的作物分类研究

时相和光谱特征信息在农作物种植分类提取方面具有十分重要的应用价值。以黑龙江大型农场--友谊农场为研究区域,利用4景不同时相的TM和SPOT卫星遥感影像,提取相应时相的NDVI时间谱图像数据作为新波段信息,在分析地物目标在相应影像各波段上光谱和时间特征的基础上,设计了决策树分类算法,通过对待分类影像进行系列阈值分割和掩膜处理,成功提取黑龙江友谊农场的大豆、玉米和水稻的种植信息,分类总体精度达到98.67%。     

基于MODIS多时相的江苏启东市油菜种植面积提取

利用遥感手段,基于油菜种植生长期的MODIS-NDVI 16d合成数据,提取江苏省启东市2008~2009、2009~2010和2010~2011年3个生长季的冬季油菜种植面积。根据油菜生育期NDVI值的时相变化建立了NDVI时序曲线;利用阈值剔除非耕地数据后,对数据利用最小噪声分离(MNF)方法进行压缩处理,最后利用波谱角分类法确定油菜种植区的空间分布,计算油菜种植面积。计算结果与实际统计面积相比,精度达到90%以上,表明该方法作为一种快速油菜种植面积监测方法,具有较高的监测精度。     

基于多时相HJ卫星的冬小麦面积提取

我国环境与灾害监测预报小卫星HJ-1A/B具有较高的时间和空间分辨率,在作物种植面积提取和长势监测等方面具有较大优势。本文以江苏省姜堰市为研究区,根据冬小麦的物候规律和季相节律的差异性,选取返青期和拔节期两个生育期的HJ卫星影像,借鉴分层信息提取法原理,综合利用监督分类和非监督分类法,结合人机交互目视解译和实地定位调查等资料提取了姜堰市的冬小麦种植面积,总体面积提取精度达到90.22%,样点空间匹配精度为81.25%,实验基地空间匹配精度为80.34%。结果表明：HJ卫星能够用于提取南方地区冬小麦种植面积和长势监测,满足农情监测的需要,且利用多时相遥感影像能有效地增加信息量,实现信息互补,有助于提高监测精度。     

应用TM图像采用模式识别技术自动提取冬小麦播种面积的研究

在农作物遥感估产研究过程中,如何快速、准确获取当年种植面积是一个关键技术问题。本文重点研究在禹城县冬小麦遥感估产试验中,应用同步TM信息源,根据冬小麦生长发育的特征,选择 TM 的适宜时相,构建多维绿度图,采用模式识别技术,分层自动提取纯麦地、套种麦地信息。这项研究结果与1/5万比例尺 TM 图像目视解译小麦面积相比较,其相对误差甚小,达到了估产实际应用的精度。     

利用RefineNet模型提取冬小麦种植信息的方法

冬小麦是我国主要的粮食作物，获取精细的冬小麦种植信息对于指导农业生产具有重要的意义。通过对RefineNet模型进行扩展，形成了适宜提取冬小麦种植信息的Ex-RefineNet（Extend-RefineNet）模型，Ex-RefineNet模型由两个子模型组成，Ex-RefineNet-Edge子模型用于提取冬小麦种植区域的边缘像素，Ex-RefineNet-Inner子模型用于提取冬小麦种植区域的内部像素，使用贝叶斯模型对两个子模型的提取结果进行合并处理，形成最终提取结果。利用山东省济南市和泰安市的16幅高分2号遥感影像进行实验，将每幅影像的2/3作为训练数据，其他数据作为测试数据，选择平均精度、查全率和Kappa系数作为对比指标，Ex-RefineNet模型的结果分别为0.93、0.92、0.91，而RefineNet模型的结果分别为0.86、0.84、0.83，说明本文给出的方法在提取冬小麦种植信息方面具有较明显的优势。     

利用RefineNet模型提取冬小麦种植信息的方法

冬小麦是我国主要的粮食作物，获取精细的冬小麦种植信息对于指导农业生产具有重要的意义。通过对RefineNet模型进行扩展，形成了适宜提取冬小麦种植信息的Ex-RefineNet（Extend-RefineNet）模型，Ex-RefineNet模型由两个子模型组成，Ex-RefineNet-Edge子模型用于提取冬小麦种植区域的边缘像素，Ex-RefineNet-Inner子模型用于提取冬小麦种植区域的内部像素，使用贝叶斯模型对两个子模型的提取结果进行合并处理，形成最终提取结果。利用山东省济南市和泰安市的16幅高分2号遥感影像进行实验，将每幅影像的2/3作为训练数据，其他数据作为测试数据，选择平均精度、查全率和Kappa系数作为对比指标，Ex-RefineNet模型的结果分别为0.93、0.92、0.91，而RefineNet模型的结果分别为0.86、0.84、0.83，说明本文给出的方法在提取冬小麦种植信息方面具有较明显的优势。     

Early Estimation of Winter Wheat Planting Area in Qingyang City by Decision Tree and Pixel UnmixingMethods based on GF-1 Satellite Data

In Northwest China,there are many mixed pixels in the winter wheat area,so the combination of decision tree and mixed pixel decomposition is of great significance to improve the interpretation accuracy.The data source of this result is GF-1 satellite data which excellent in the high temporal resolution and high spatial resolution.Based on the difference about variation characteristics and NDVI value for winter wheat and the other crops in different phase data,we build decision tree to extract winter wheat pixels preliminary.Then selected linear spectral mixture model,further analysis the previous data by mixed pixel decomposition,get the final planting area data more exactly.Compared with the winter wheat samples measurement data,calculate the extraction accuracy eventually.The result shows that the extraction accuracy of winter wheat planting area in the study area was more than 90%,Kappa coefficient is close to 0.8,can reflect the distribution of winter wheat in the region accurately.This study found that the method which combined with decision tree classification and pixel unmixing based on high resolution remote sensing image can extract the winter wheat planting area precisely,This is helpful for the development of crop area remote sensing monitoring.     

PlantingInformation Extraction of Winter Wheatand Rape based on GF-1 Images

Successive emission of high resolution satellite has created new opportunities for the application of domestic high resolution remote sensing data.In order to explore the feasibility of GF data in the field of small and medium scale crop remote sensing monitoring and to establish a suitable technical system,with Yangzhou as an example,using decision tree model and object oriented classification method to research the feasibilityon crop planting information extraction of GF wide field viewdata.And explore the method to improve the accuracy.The results showed that,sub\|regionpretreatmentcan reduce the adverse effects of crop spatial distribution on the extraction of the planting area.The overall accuracy of winter wheat was 97%,the Kappa coefficient was 0.93;the overall accuracy of rape was 96%,the Kappa coefficient was 0.84.Research shows thatdomestic GF\|1 WFV images can be applied to the crop planting informationextraction,and toprovide an important reference and decision support for adjusting crop spatial and optimizing management of gain producing areas.     

基于ISODATA法的冬小麦产量分级监测预报

以江苏省姜堰市为例,进行了基于TM卫星遥感技术和小麦估产模型的冬小麦产量监测研究。在利用GPS实地采样调查和建立解译标志的基础上,通过影像校正、采用优化的ISODATA分类方法,结合人机交互式判读解译等操作,将样点的作物信息数据贯穿到整个校验分类过程中,信息解译精度在90%以上。利用分类提取的冬小麦数据,反演叶面积指数、生物量信息等,结合冬小麦估产模型,计算单点产量信息,经过线性转换,对整个区域的冬小麦产量进行监测预报,并制作了冬小麦产量分级专题图。     

Winter wheat mapping using temporal signatures of MODIS vegetation index data

Because most land-cover types have distinct seasonal changes and corresponding reflectance characteristics in remotely sensed images, the signatures in time-series data are useful for discriminating different land covers. Although temporal signatures have been used to classify different land-cover types, they have not been fully exploited to classify specific crops, and the influence of low resolution should be evaluated. The aims of this study were to seek an effective method to classify specific crops using the temporal signatures in coarse time-series data and to examine the applicability of the data for crop classification as well. A winter wheat-producing region in China was selected for this case study. Moderate-Resolution Imaging Spectroradiometer (MODIS) 8-day composite land surface reflectance product (MOD09Q1) data with a 250 m spatial resolution were used to calculate the vegetation index data, which was applied to detect the properties of live green plants. The noise in the time series was filtered to minimize the classification uncertainties. The curve shape in the time-series vegetation index profile was used as the major metric to classify winter wheat, and other phenological metrics extracted from the data were used conjunctly as auxiliary functions to improve the separability. The metrics for winter wheat classification were quantified in the large fields with relatively pure pixels. Winter wheat was successfully extracted from the MODIS vegetation index data, and the MODIS-derived result was validated with a fine-resolution (19.5 m) thematic map derived from images collected by the charge-coupled device sensor on board the China–Brazil Earth Resources Satellite (CBERS). It showed that the MODIS-derived result had inevitable low-resolution bias, and the errors of commission and omission were 32.3 and 33.8%, respectively. The overall classification effect of the MODIS-derived result relied upon the distribution of pixel purity in the study area.     

基于WorldView-2影像和语义分割模型的小麦分类提取

为使用高分辨率遥感影像和深度学习语义分割模型实现快速准确的小麦种植空间信息提取,以WorldView-2遥感影像为数据源,制作尺度分别为128×128、256×256、512×512的样本数据集,对U-net和DeepLab3+语义分割模型的参数进行训练,建立小麦遥感分类模型;通过与极大似然和随机森林方法比较,检验深度学习分类效果。结果显示：(1)不同尺度样本训练得到的模型总体精度、Kappa系数分别在94%和0.82以上,模型精度稳定,样本尺度大小对小麦分类提取模型影响较小;(2)深度学习方法的小麦分类总精度和Kappa系数分别在94%和0.89以上,极大似然和随机森林则在92%和0.85以下,表明该研究建立的小麦遥感分类模型优于传统分类方法。研究结果可为高分辨率遥感影像作物种植信息的深度学习方法提取提供参考。     

The area extraction of winter wheat in mixed planting area based on Sentinel-2 a remote sensing satellite images

ABSTRACT

The traditional area extraction method mainly depends on manual field survey methods, it is workload, slow and high cost. While remote sensing technology has the advantages of accuracy, rapidity, macroscopic and dynamic, which has become an effective means to extract crop growing area. In this paper, we took Kaifeng City in Henan Province as the study area. Firstly, we explored the advantages of Sentinel-2A RENDVI in crop identification. Then used the supervised classification SVM, object-oriented classification method and assisted with field measured data to extract the winter wheat planting area, the characteristics of the two methods were compared and analysed. Finally, we combined the above two classification methods and proposed a new classification method V2OAE to remove unnecessary influencing factors. The experiment results showed that RENDVI has better recognition ability than the NDVI (Normalized Difference Vegetation Index) in distinguishing vegetation with similar spectrum, the classification effect of object-oriented classification is better than supervised classification SVM, and our classification method removes unnecessary influence factors in the results of object-oriented classification, which is further improve the monitoring accuracy.

Firstly, we have preprocessed the Sentinel-2A image data, its steps are: (1) In the first step, we made radiation calibration for remote sensing images to eliminate the image distortion caused by external factors, data acquisition and transmission systems and so on; (2) In the second step, we made atmospheric correction to eliminate changes in the spectral feature of remote sensing images caused by atmospheric absorption or scattering; (3) In the third step, we made band resampling to unify the resolution of remote sensing images and facilitate the mathematical combination operation of vegetation index; (4) In the fourth step, we made mosaic and cutting to get preprocessed remote sensing images of Kaifeng City. Secondly, we analysed the spectral features of each object and established the interpretation mark with the field measured data. then we explored the ability to identify the ground objects based on NDVI(Normalized Difference Vegetation Index) and RENDVI. Third, we used the rule-based object-oriented classification method and SVM classification to extract the planting area of the study area, the input definition of SVM is spectral feature images of ground objects and the output definition of SVM is the recognition result of ground objects in the process of data training. Then the advantages and disadvantages of the two methods in classification results were analysed. Finally, In order to extract winter wheat information more accurately, we combined the above two classification methods and proposed a new classification method V2OAE (Vector Object Oriented Area Extraction) to remove unnecessary influencing factors, then the winter wheat planting area in Kaifeng City was statistically obtained.     

基于多时相TM影像的冬小麦面积变化监测

利用北京1992年、2000年、2004年、2009年的多时相Landsat TM5影像数据,结合实际调查数据,分析了近20年来北京冬小麦种植面积的变化趋势及演变特征。采用决策树、PCA、缨帽变换等手段对地物进行分类,利用多时相影像,NDVI组合阈值提取小麦种植区面积。研究结果表明:北京地区1992年、2000年、2004年、2009年冬小麦种植面积分别为:113671ha,84322ha,40410ha,61529ha。北京冬小麦种植面积呈现为明显的先减少后增加的趋势。从1992年到2009年共减少52143ha。其中,从1992年到2000年冬小麦种植面积减少了29349ha,减少的面积中城区扩张占用和转变为裸地的最多,分别为39.7%和42.8%,另外有13.3%变为设施用地,3%成为水体(鱼塘和水田);从2000年到2004年冬小麦种植面积共减少43921ha,减少的面积中转变为裸地和城区扩张占用的最多,分别为39.8%和33.1%;从2004年到2009年冬小麦种植面积共增加了21119ha,其中裸地转变为小麦种植区面积最大。     

利用航空成像光谱数据进行冬小麦产量预测

以国产成像光谱仪PHI(Pushbroom Hyperspectral Imaget)所获遥感影像数据为基础，根据田间冬小麦单产遥感研究试验数据建立了研究区不同时相冬小麦单产预测模型，实现了利用航空高光谱遥感数据对研究区小麦产量的整体预测；对试验区土壤氮素水平与不同时相冬小麦预测产量以及试验区实测产量进行了初步分析，分析结果显示：土壤氮素分布的差异性对小麦的产量有明显影响。