Caracterización del efecto de estrés usando índices espectrales de la vegetación para la estimación de variables relacionadas con la biomasa del área (original) (raw)
Related papers
Review of spectral vegetation indices and methods for estimation of crop biophysical variables
Aerospace Research in Bulgaria, 2017
In present article a brief overview is presented on spectral vegetation indices and methods for estimation of crop main biophysical variables and their proxies. The main VIs used in estimation of nitrogen and chlorophyll, biomass, LAI and fAPAR, fCover, and photosynthesis are summarized. Biophysical variables and vegetation indices A number of techniques have evolved to derive the biophysical variables of vegetation using remote sensing data; these can be grouped into three broad categories: the inversion of radiative transfer models [39], machine learning (for example neural networks) [4] and the use of vegetation Indices. There are generally few ways of deriving the biophysical estimates using empirical or semi-empirical relationships: 1) single regression; 2) stepwise linear regression; 3) partial least squares (PLS) regression; 4) artificial neural networks [12]. Methods based on vegetation indices (VIs) have the benefit of being computationally simple while they are generally less site specific and more universally applicable than the other methods. The performance of the different indices and selected "optimal" wavebands depends on vegetation and land cover type, the variables to be retrieved, sun/view geometry to name but a few [12]. Satellite spectral data has the potential to measure the reflected radiation from many plants, thus making assessment of biophysical variables feasible on canopy level. The regression models relate in situ measurements and VIs. The VIs are mathematical transformations of the original spectral reflectance that are designed to reduce the additive and multiplicative errors associated with atmospheric effects, solar illumination, soil background effects, and sensor viewing geometry [29].
Spectral Data for Determination of Crop Vegetation Indices
The field experiments with four levels of N-fertilizer (0, 30, 60, and 90 kgℎ −1 ) in two repetitions were conducted for three years to select some appropriate vegetation indices for winter wheat. Hyper-spectral reflectance data using a portable field spectroradiometer (351 to 2,500 nm) were recorded from 10 am to 2 pm under cloudless conditions at two different growth stages of winter wheat. All two-band and three band combinations of several vegetation indices were subsequently calculated in an algorithm for determining linear regression analysis against SPAD value, protein content, and grain yield. R square matrices were used to make contour plots and 3-D scatters. Using overlaying in analysis tools of ArcMap the between first and second year results, a number of common hot spots with strong correlations were revealed. The selected hot spots were validated with the dataset of the third year to choose the best vegetation indices for crop variable estimations.
A multi-temporal comparison was made of four spectral vegetation indexes among 50 sites in a subtropical, coastal arid region. Greenness indexes (NDVI, SAVI and TSAVI) and one of moisture (NDII) were evaluated. NDVI and SAVI were very closely correlated, while TSAVI fluctuated less and NDII showed strong seasonal variations. Topographic correction (illuminated surface) of raw data usually increased the slope of TSAVI's soil line by >20%. The indexes, except NDII, were used to estimate Leaf Area Index; signal-to-noise analysis of LAI suggested that SAVI is usually close to NDVI but TSAVI has much more signal in the drier months.
Vegetation Stress Indicators Derived from Multispectral and Multitemporal Data
Remote sensing is already an operational tool widely used in vegetation studies for ecological monitoring, change detection of natural ecosystems and in agriculture for crop state assessment and yield prediction. A strong stress is being put on the accuracy of the retrieved information. This requires reliable indicators of plant growth and physio-logical status. The development of efficient means for data analysis is still one of the most essential issues. The importance of this issue is directly related to the ever-increasing amount of data provided by numerous sensors. The use of multi-spectral and multitemporal remotely sensed data and the implementation of advanced data processing technologies results in the possibility of getting different information needed for decision-making in solving problems related to vegetation preservation and agricultural land use. The application of satellite data requires knowledge of land covers spectral behaviour under different environmental cond...
DISEÑO DE UN ÍNDICE ESPECTRAL DE LA VEGETACIÓN: NDVIcp DESIGN OF A VEGETATION SPECTRAL INDEX: NDVIcp
2007
There are many vegetation indices (VI) based on relationships of the spectral space of the red and near infrared. In this study, the structure of the most widely used VI is examined, using a formulation to characterize curves of equal leaf area index. In order to solve the inconsistencies found in the VI, a new one (NDVIcp) is proposed, based on the correct structure of the problem, under empirical considerations. The NDVIcp is validated using data from field experiments with maize (Zea mays L.) and cotton (Gossipyum spp.).
2013
Early crop yield estimation is extremely relevant to assist on management decisions and to plan harvest logistics and commercialization. The purpose of this study is to evaluate the ability of widely reported spectral indices to estimate leaf area index (LAI) and total above ground biomass (DW) through the season, and their capacity to provide useful information for crop yield estimation. Particular emphasis is placed in identifying indices that can estimate the biophysical parameters of crops that accumulate large amounts of DW and LAI. The study site located in the south west of Uruguay was planted with wheat cultivar Biointa1006. From emergence until maturity the site was imaged approximately every 15-20 days with a UAV carrying a multispectral camera with channels in the green, red and NIR regions. Three indices were estimated, NDVI, NDVIgreen, CIgreen. Approximately at the same date of the flight sampling points were visited, DW and LAI were measured. The evolution of indices r...
Desarrollo de curvas espectrales del crecimiento anual de la vegetación, usando sensores remotos
Revista Mexicana de Ciencias Pecuarias, 2011
Development of historic vegetation growth curves for both grasslands and shrublands in Mexico becomes a real challenge due to lack of statistics. Remote sensing technology using satellites, allows obtaining a first approximation through vegetation indices (VI). The present study discusses an indirect parametric modeling of the biomass growth curve in individual pixels of both Advanced Very High Resolution Radiometer (AVHRR) and Moderate Resolution Imaging Spectroradiometer (MODIS), thus allowing their characterization using pre and post process algorithms. Results obtained through application of the algorithms discussed in the present paper, allow obtaining annual growth curves at 100 ha (AHVRR) and 6.25 ha (MODIS) scales that can be used in studies on climate change and on grassland/shrubland management in Mexico.
Applications of Vegetative Indices from Remote Sensing to Agriculture: Past and Future
Inventions
Remote sensing offers the capability of observing an object without being in contact with the object. Throughout the recent history of agriculture, researchers have observed that different wavelengths of light are reflected differently by plant leaves or canopies and that these differences could be used to determine plant biophysical characteristics, e.g., leaf chlorophyll, plant biomass, leaf area, phenological development, type of plant, photosynthetic activity, or amount of ground cover. These reflectance differences could also extend to the soil to determine topsoil properties. The objective of this review is to evaluate how past research can prepare us to utilize remote sensing more effectively in future applications. To estimate plant characteristics, combinations of wavebands may be placed into a vegetative index (VI), i.e., combinations of wavebands related to a specific biophysical characteristic. These VIs can express differences in plant response to their soil, meteorolog...