Evaluation and analysis of training sets and training statistics of multispectral scanner data related to agricultural crop cover types

The size and reliability of the training sets or sample area for the classification of airborne multispectral scanner data obtained over an agricultural area with the help of an interactive computer system have been examined in this study. The experiment reported herein suggests that a training set of not less than 50 pixels would adequately represent all the likely variations in any particular field. The evaluation of the results further reveals that if the training sets can adequately represent the field variations characteristic of the region, the corresponding training statistics can be utilized both on scanline and pixel directions.     

光谱特征扩展的时间序列Landsat数据地表覆盖分类

时间序列遥感影像常用于地表覆盖监测及其变化监测。然而,利用时序遥感数据—尤其是中分辨率遥感数据监测地表覆盖变化,其方法基本是先对多期影像分别进行监督分类然后对比分类结果。由于这种方法需要对每期遥感影像单独选择分类训练样本,而对于历史影像,常常难以获得可靠的样本数据。本文基于遥感数据定量化处理,尝试利用光谱特征扩展方法对时间序列Landsat数据进行分类:首先,结合一种新的大气校正方法和相对辐射归一化方法,对时间序列Landsat数据进行定量化处理,以消除各期影像之间的辐射差异,获得地表反射率数据。然后,论文选择一期易于获得分类训练样本的反射率数据作为"参考影像",并结合样本数据提取不同地表覆盖类型的光谱特征。最后,将"参考影像"中提取的地物光谱特征,扩展到所有时间序列反射率数据进行分类。论文利用青藏高原玛多地区的5景Landsat数据对本文的方法进行了验证,结果显示:基于光谱特征扩展的分类方法,可有效对定量化处理后的Landsat数据进行分类,分类总体精度为88.35%—94.25%,分类结果和传统的单景监督分类结果具有较好的一致性。此外,研究也发现,"参考影像"和待分类图像获取时间的季相差异会影响其分类的精度。     

Comparison of ERS-SAR and spot data for sugar beet crop cover assessment

The present work was aimed to compare the abilities of radar and optical satellite data to estimate crop canopy cover, which is a key component of productivity estimates. Three ERS-1 SAR images were obtained of East Anglia (UK) in 1995 and one ERS-2 SAR image in 1996. The images covered a study area around the IACR Brooms Barn Sugar Beet Research Institute. Field data comprising radiometric and biophysical measurements of the crop canopy were collected in two fields from June 22 to August 3, 1995 to coincide with ERS-1 SAR overpass dates. In 1996, field data were collected in two fields from June 11 to July 29 on a weekly basis. A previously calibrated version of the water cloud model was inverted to estimate Leaf Area Index (LAI) from ERS-1 and ERS-2 SAR backscatter and soil moisture samples. Canopy cover was estimated from the radar-estimated LAI using a standard exponential relationship that has a well-established coefficient for sugar beet. Radio-metrically and atmospherically corrected data from three SPOT images in 1995 and one SPOT image in 1996 were used to calculate the Optimised Soil Adjusted Vegetation Index (OSAVI), from which crop canopy cover was estimated using a relationship determined previously by canopy modelling. The crop cover values estimated by satellite were in good agreement with those measured on ground with the Parkinson radiometer. Radar data may be able to provide useful estimates of canopy cover for crop production modelling, especially in the case of loss of optical data due to cloud.     

Crop cover and crop phenological information from red and infrared spectral responses

Possibility of utilizing the red and infrared spectral information for assessing status of vegetation cover and consequential crop phenological information are discussed. The experiment was conducted in a potential agricultural area around Mandya town of Karnataka State and airborne spectral information was obtained through modular multispectral scanner from a height of 1000 meters above the ground level. The spectral information of red (0.66–0.70 urn) and infrared (0.77–0.86 urn) bands was extracted with the aid of an interactive computer system : the multispectral data analysis system. Based on the spectral information, the data was analysed and interpreted with the support of ground information. Crop fields without vegetation were observed to have infrared/red ratio in the range of 0.70 to 0.97 and also it was possible to distinguish wet and dry paddy field. Crop fields covered with vegetation exhibited higher infrared/red ratio depending on the nature of crop growth. For instance, rice crop exhibited spectral ratio of 0.78 at the time of planting, 3.52 at the time of maximum vegetation growth and 2.04 during the maturation phase. In case of sugarcane crop, the increase and decrease in spectral ratio were gradual because of its longer duration. From infrared and red band information it was possible to distinguish crop species based on rate of change of vegetation cover which corresponded with the change in spectral ratios. The temporal information expressed in two dimensional space for red and infrared band also enabled clearly to distinguish between rice and sugarcane.     

农作物冠层光谱分析及反演技术综述

农作物的冠层光谱反射率与作物的氮含量、叶绿素含量及叶面积指数等参数之间具有很强的相关性,通过对作物冠层光谱进行分析可反演出作物的生物物理参数,并应用在长势分析、产量预测、病虫害预警等领域。本文首先阐述了作物冠层反射率采集方法,对地面、机载及遥感卫星3个采集层面的优缺点进行了对比;其次给出了植被指数构建原理及常用植被指数,分析了物理模型反演法和统计反演法的复杂度和性能;最后提出了农作物冠层光谱分析及反演技术的下一步发展方向及面临的挑战。     

Effect of solar elevation on rice crop spectral response

In the present study the effect of solar elevation angle on the spectral response of rice crop was examined under farmer’s field conditions. The aim of the study is to see if band ratioing can reduce the effect of soiar elevation angle on rice crop spectral response. It was observed that the spectral transformations involving red and near infrared are highly useful in normalizing the effect of sun elevation when the canopy cover is complete. Contrary to this, the spectral transformation could not normalize the effect of sun elevation on the spectral response of rice crop when the canopy cover is incomplete.     

Contrast enhancement of landsat data on multispectral data analysis system

Contrast enhancement, one of the image processing techiques, is developed on the Multispectral Data Analysis System (MDAS) for enhancing the LANDSAT data. The purpose of image processing for enhancement is to improve the obscure objects data in the image to stand out more readily for good sensing to the human eye. It is observed on MDAS that some of the LANDSAT scenes when examined on the color display, give inadequate information for the required objective of interpretation. This is due to poor tonal contrast in the scene because of prevailing climatological conditions at the time of satellite pass over that area. Also, the LANDSAT data usually occupy a small subset of the total brightness range 0–127. To provide optimal contrast and variation for color compositing, contrast enhancement may by performed on the data before going to trie information processing (categolization) on the landsat scene. This paper describes the algorithms of parametric linear and non linear contrast enhancement techniques. A typical example to differentiate the degree of salinity in the soils was tested with the suggested algorithms and the results are tabulated in the form of photographs. The test area is selected from Haryana (frame no. 158-040 dated 2nd May, 1977) for testing the algorithms. The enhancement software developed on the MDAS stretches all the four Landsat bands and generates an output tape with the format similar to LANDSAT computer compatible tape (CCT). The stretched results of 5 and 7 bands are displayed in this paper. A false color composite which appears as on the color displya could also be generated from 4, 5 and 7 bands. The enhanced output was found to be useful for easily categorizing the data into various categories on MDAS.     

Use of spectral data in Markov chain model for crop yield forecasting

Crop yield is mainly dependent on weather, soil and technological inputs. Yield forecasting models have been developed mainly using multiple regression techniques based on biometrical characters of the plants and/or weather parameters. Matiset al. (1985) proposed another approach of crop yield modelling using Markov Chain theory based on biometrical characters. The integration of remote sensing with other technologies has provided an immense scope to improve upon the existing crop yield models. In the present study, multi date spectral data during crop growth period was used in Markov Chain Model to forecast wheat yield. The results indicate that the use of spectral data near the maximum vegetative growth of wheat crop improves the efficiency and reliability of yield forecast about a month before its actual harvest.     

机载Lidar数据的农作物覆盖度及LAI反演

虽然Lidar点云数据已被广泛应用于获取森林各项结构参数,但这些方法并不适合于低矮的灌丛、林地和农作物。本文以玉米为研究对象,提出利用机载Lidar点云数据的强度信息和全波形数据中的距离与扫描天顶角信息,反演农作物覆盖度和LAI的方法。在黑河进行的飞行实验和地面验证表明,该方法具有较高精度,也表明Lidar在低矮自然植被监测和农业应用上有较大潜力。     

Study of fractional vegetation cover using high spectral resolution data

A field experiment was conducted to study the effect of vegetation cover on soil spectra and relationship of spectral indices with vegetation cover. Multi-date spectral measurements were carried out on twelve wheat fields. Five sets of measurements were taken during the growth period of wheat crop. Field reflectance data were collected in the range 350 to 1800 nm using ASD spectroradiometer. Analysis of data was done to select narrow spectral bands for estimation of ground cover. The ratio of reflectance from vegetation covered soil and reflectance from bare soil indicated that spectral reflectance at 670 and 710 nm are the most sensitive bands. Two bands in visible (670 and 560 nm), three bands in near infrared (710, 870 and 1100 nm) and three bands in middle infrared (1480, 1700 and 1800 nm) were found highly correlated with fractional cover. Vegetation indices developed using narrow band spectral data have been found to be better than those developed using broad- band data for estimation of ground cover.     

An integrated methodology for soil moisture analysis using multispectral data in Mongolia

Soil moisture (SM) content is one of the most important environmental variables in relation to land surface climatology, hydrology, and ecology. Long-term SM data-sets on a regional scale provide reasonable information about climate change and global warming specific regions. The aim of this research work is to develop an integrated methodology for SM of kastanozems soils using multispectral satellite data. The study area is Tuv (48°40′30″N and 106°15′55″E) province in the forest steppe zones in Mongolia. In addition to this, land surface temperature (LST) and normalized difference vegetation index (NDVI) from Landsat satellite images were integrated for the assessment. Furthermore, we used a digital elevation model (DEM) from ASTER satellite image with 30-m resolution. Aspect and slope maps were derived from this DEM. The soil moisture index (SMI) was obtained using spectral information from Landsat satellite data. We used regression analysis to develop the model. The model shows how SMI from satellite depends on LST, NDVI, DEM, Slope, and Aspect in the agricultural area. The results of the model were correlated with the ground SM data in Tuv province. The results indicate that there is a good agreement between output SM and SM of ground truth for agricultural area. Further research is focused on moisture mapping for different natural zones in Mongolia. The innovative part of this research is to estimate SM using drivers which are vegetation, land surface temperature, elevation, aspect, and slope in the forested steppe area. This integrative methodology can be applied for different regions with forest and desert steppe zones.     

基于谱间特征与多种指数分析的居民地信息提取方法

本文在分析现有居民地提取方法的基础上,提出将归一化建筑指数(NDBI)、改进归一化差异水体指数(MNDWI)、土壤调节植被指数(SAVI)、比值居民地指数(RRI)相结合进行居民地信息提取的方法。以浙江省宁波市为例,通过光谱采样及各类地物在4种指数上的取值分析,建立模型进行居民地信息提取及精度验证,结果表明:利用该模型可以实现居民地信息的自动提取,能提高居民地与裸地的可分性,减少背景地物的影响,总体精度为91.08%。     

Estimating fractional cover of crop,crop residue,and soil in cropland using broadband remote sensing data and machine learning

The fractional vegetation cover (FVC), crop residue cover (CRC), and bare soil (BS) are three important parameters in vegetation–soil ecosystems, and their correct and timely estimation can improve crop monitoring and environmental monitoring. The triangular space method uses one CRC index and one vegetation index to create a triangular space in which the three vertices represent pure vegetation, crop residue, and bare soil. Subsequently, the CRC, FVC, and BS of mixed remote sensing pixels can be distinguished by their spatial locations in the triangular space. However, soil moisture and crop-residue moisture (SM-CRM) significantly reduce the performance of broadband remote sensing CRC indices and can thus decrease the accuracy of the remote estimation and mapping of CRC, FVC, and BS. This study evaluated the use of broadband remote sensing, the triangular space method, and the random forest (RF) technique to estimate and map the FVC, CRC, and BS of cropland in which SM-CRM changes dramatically. A spectral dataset was obtained using: (1) from a field-based experiment with a field spectrometer; and (2) from a laboratory-based simulation that included four distinct soil types, three types of crop residue (winter-wheat, maize, and rice), one crop (winter wheat), and varying SM-CRM. We trained an RF model [designated the broadband crop-residue index from random forest (CRRF)] that can magnify spectral features of crop residue and soil by using the broadband remote sensing angle indices as input, and uses a moisture-resistant hyperspectral index as the target. The effects of moisture on crop residue and soil were minimized by using the broadband CRRF. Then, the CRRF-NDVI triangular space method was used to estimate and map CRC, FVC, and BS. Our method was validated by using both laboratory- and field-based experiments and Sentinel-2 broadband remote-sensing images. Our results indicate that the CRRF-NDVI triangular space method can reduce the effect of moisture on the broadband remote-sensing of CRC, and may also help to obtain laboratory and field CRC, FVC, and BS. Thus, the proposed method has great potential for application to croplands in which the SM-CRM content changes dramatically.     

Wheat crop inventory using high spectral resolution irs-p3 mos-b spectrometer data

Modular Optoelectronic Scanner (MOS-B) spectrometer data over parts of Northern India was evaluated for wheat crop monitoring involving (a) sub pixel wheat fractional area estimation using spectral unmixing approach and (b) growth assessment by red edge shift at different phenological stages. Red shift of 10 nm was observed between crown root initiation stage to flowering stage. Wheat fraction estimates using linear spectral unmixing on Feb. 13, 1999 acquisition of MOS-B data had high correlation (0.82) with estimates from Wide Field Sensor (WiFS) data acquired on same date by IRS-P3 platform. It was observed that five bands (4,5,8,12,13 MOS-B bands) are sufficient for signature separability of major land cover classes viz. wheat, urban, wasteland, and water based on purely spectral separability criterion using Transformed Divergence (T.D.) approach. Higher number of bands saturated the T.D. values. In contrast, performance of sub pixel fractional area estimation using unmixing decreased drastically for eight bands (4,5,6,7,8,9,12,13 MOS-B bands) chosen from optimal band selection criteria in comparison to full set of 13 bands. The relative deviation between area estimated from Wifs and MOS-B increased from 1.72 percent when all thirteen bands were used in unmixing to 26.10 percent for the above eight bands.     

Data structure characterization of multispectral data using principal component and principal factor analysis

Abstract

Both principal component analysis (PCA) and principal factor analysis (PFA) were used to analyze an experimental multispectral data structure in terms of common and unique variance. Only the common variance of the multispectral data was associated with the principal factor, while higher‐order principal components were associated with both common and unique variance. The unique variance was found to represent small spectral variations within each cover type as well as noise vectors, and was most abundant in the lower‐order principal components. The lower‐order principal components can be useful in research designed to discriminate minor physical variations within features, and to highlight localized change when using multitemporal‐multispectral data. Conversely, PFA of the multispectral data provided an insight into a great potential for discriminating basic land‐cover types by excluding the unique variance which was related to the noise and minor spectral variations.     

结合Sentinel-2光谱与纹理信息的冬小麦作物茬覆盖度估算

作物茬覆盖度的估算对于探究农业耕作方式对周围环境的影响具有十分重要的意义。目前,基于多光谱影像的作物茬指数是作物茬覆盖度估算的常用方法。然而,在作物茬高覆盖区域,指数法容易出现“饱和”现象。已有研究结果表明结合影像的光谱与纹理信息有助于改善指数法的“饱和”问题。Sentinel-2作为一颗多光谱卫星,空间分辨率可达10 m,与Landsat OLI相比,具有更丰富的纹理信息。因此,探究Sentinel-2光谱与纹理信息相结合在作物茬覆盖度估算上的潜力具有重要意义。本文以山东省禹城市为研究区,分析了Sentinel-2各波段反射率、归一化差值指数以及不同窗口大小下灰度共生矩阵统计量等遥感因子与野外实测作物茬覆盖度的相关性,并利用最优子集法对遥感因子进行筛选,构建作物茬覆盖度的最优估算模型。同时,使用留一法交叉验证对模型进行评价。结果表明在单因子分析中,归一化差异耕作指数NDTI（Normalized Difference Residue Index）与作物茬覆盖度的相关性最好,相关系数达0.735。使用NDTI、5×5窗口下Sentinel-2 8A波段的相关性统计量以及12波段的方差统计量构建的多元方程是作物茬覆盖度估算的最优模型,相关系数为0.869,均方根误差为11%。与仅使用光谱信息的最优模型相比,相关系数提高了0.094,均方根误差下降了3.5%。可见,结合Sentinel-2的纹理信息有助于提高作物茬覆盖度的估算精度。     

Hierarchical clustering of multispectral images using combined spectral and spatial criteria

An agglomerative hierarchical clustering method, which uses both spectral and spatial information for the aggregation decision, is proposed here. The method is suitable for large multispectral images, provided that an unsupervised classification is previously applied. The method is tested on a synthetic image and on a satellite image of the coastal zone.     

出租车数据的时间序列谱聚类分析

城市空间运行的出租车产生大量的OD数据,数据的时序呈现周期性特点,客观反映人们的出行行为模式,本文采用谱聚类算法对北京五环区域内各空间单元的出行特征及其相似性进行探究。由于空间单元的时空行为特征受空间邻域和功能区划的影响,研究添加邻域因子和功能区因子以改进时间序列的相似性度量方法,从而实现时间序列谱聚类算法的空间和功能区拓展,进而增加人们对不同时空条件下出行行为特征的了解,以便对不同空间单元提供差异性服务,如不同时段公交的发车频次、动态调整商场营业时间、不同时空环境出租车候车点的实时变换、调控和优化不同区域服务保障等,将有助于降低城市能耗,更加合理地利用资源,也有助于居民实现智慧生活。     

Assessment of crop growth parameters of wheat under stress condition through ground based spectral data

A field experiment was conducted on wheat crop during rabi seasons of 1995–96, 1996–97 and 1997–98 to study the spectral response of wheat crop (between 490 to 1080 nm) under water and nutrient stress condition. An indigenously developed ground truth radiometer having narrow band in visible and near infrared region (490 – 1080 nm) was used. Vegetation indices derived using different band combinations and related to crop growth parameters. The near infrared spectral region of 710 – 1025 nm was found most important for monitoring stress condition. Relationship has been developed between crop growth parameters and vegetation indices. Leaf Area Index (LAI) and chlorophyll could be predicted by knowing different reflectance ratios at milking stage of crop with R² value of 0.78 and 0.89, respectively. Dry biomass (DBM), Plant Water Content (PWC) and grain yield are also significantly related with reflectance ratios at flowering stage of crop with R² value of 0.90, 0.98 and 0.74, respectively.     

Testing the capability of spectral resolution of the new multispectral sensors on detecting the severity of grey leaf spot disease in maize crop

Abstract

In this study, we tested whether GLS field symptoms on maize can be detected using hyperspectral data re-sampled to WorldView-2, Quickbird, RapidEye and Sentinel-2 resolutions. To achieve this objective, Random Forest algorithm was used to classify the 2013 re-sampled spectra to represent the three identified disease severity categories. Results showed that Sentinel-2, with 13 spectral bands, achieved the highest overall accuracy and kappa value of 84% and 0.76, respectively, while the WorldView-2, with eight spectral bands, yielded the second highest overall accuracy and kappa value of 82% and 0.73, respectively. Results also showed that the 705 and 710 nm red edge bands were the most valuable in detecting the GLS for Sentinel-2 and RapidEye, respectively. On the re-sampled WorldView 2 and Quickbird sensor resolutions, the respective 608 and 660 nm in the yellow and red bands were identified as the most valuable for discriminating all categories of infection.