Hyperspectral Imaging and Stress Mapping in Agriculture: A Case Study on Wheat in Beauce (France)

doi:10.1016/S0034-4257(98)00049-2

Remote Sensing of Environment

Volume 66, Issue 2, November 1998, Pages 179-191

https://doi.org/10.1016/S0034-4257(98)00049-2 Get rights and content

Abstract

In an attempt to extract information relevant for agriculture in remotely sensed wheat crops, MIVIS hyperspectral images are analyzed in the visible and near-infrared domains. Through the selection, by means of a principal component analysis (PCA), of two endmembers of wheat, related respectively to well-developed and stressed plants, a water deficiency is detected among the spectral population of wheat. The image is then modeled by a spectral mixture analysis (unmixing) of these two wheat endmembers, soil, and shade. Resulting fraction images are interpreted in terms of crop vitality (level of green biomass) in relation to stress presence and compared to field knowledge. In addition, these images allow mapping the leaf area index (LAI) over the whole scene, with an empirical relationship based on 12 ground measurements of this variable. This work shows the interest of the approach combining PCA and unmixing for stress detection and mapping of agronomic variables, with a good accuracy compared to spectral ratio analysis. It provides relevant support for crop monitoring and precision agriculture, by means of numerical cartographic products obtained by hyper- (super-) spectral remote sensing. It demonstrates the need for improved methodologies derived from hyperspectral data analysis, and reveals that, through such methods, one can, however, retrieve a significant amount of information with limited number of spectral channels (10–20), highlighting the potential of superspectral observations.

Introduction

In the context of a growing interest for remote sensing applied to vegetation studies [see, e.g., reviews by Verstraete et al. (1996) or Moran et al. (1997); and references therein], we investigate the possibilities of hyperspectral imaging observations for the extraction of information relevant to agriculture.

Such data have already proven to be relevant to many requirements of agricultural monitoring. These include, for instance, the mapping of vegetation species and classifications Foody and Cox 1994, Clark et al. 1995, Erol and Akdeniz 1996, Grignetti et al. 1997, vegetation coverage or density Huete 1986, Roberts et al. 1993, Jasinski 1996, Pax-Lenney and Woodcock 1997, or crop yield predictions (MacDonald and Hall, 1980; Clevers, 1997). But only a few of these studies (e.g., Clark et al. 1995, Clevers 1997) have addressed crop growth and development using the mapping aspect provided by imagery.

In this article we explore, in a case study of wheat fields in France, some methods that are able to trace crop evolution among several cultivated parcels and to provide the spatial distribution of parameters relevant to agricultural studies.

Section snippets

Airborne Hyperspectral Data

Hyperspectral data were acquired by the Daedalus Multispectral Infrared and Visible Imaging Spectrometer (MIVIS), flown by the Consiglio Nazionale delle Ricerche (CNR) of Italy, during a campaign in the Beauce region of France in 1996. The present work concerns an agricultural area (about 2×4 km, cf. Fig. 1) managed by the French Institut Technique des Céréales et des Fourrages (ITCF), in the vicinity of Boigneville, on 29 May.

Measurements of the first spectrometer (“optical port 1”: visible

Data analysis

In the aim of extracting some relevant information from the remotely sensed data, we have linked two methods in current use in hyperspectral analysis, in a coupled approach. First, a principal component analysis (cf., e.g., Johnson and Wichern, 1982), or PCA, leads to an understanding of the spectral variability contained in the data and of its spatial organization in the scene of study. It allows us to determine the main spectral trends and the extreme spectral classes (endmembers) which best

Stress Mapping

To understand more precisely the meaning of the fractions, the observed spectra of several pixels representing varying W1 and W2 fractions are plotted in Figure 5, for different ranges of soil fraction (0–5%, 5–10%, 10–15%, >15%). It shows that the spectrum varies continuously, with a shallowing of the absorption band in the range 0.54–0.70 μm and a darkening at greater wavelengths, as W1 fraction value decreases. This behavior, which is globally related to a continuous decrease of the

Conclusion

This study shows the capability of linear spectral mixture modeling, combined with a principal component analysis, to access relevant agricultural or agronomic information in a natural scene. This analysis leads to the decomposition of the detected spectrum into several constituents, and allows us to discard soil and shade contributions from the vegetation spectrum. The method also provides a relevant tool for the detection of the presence of some stresses (e.g., hydrous and nitric), giving the

Acknowledgements

The campaign of data acquirements with the MIVIS instrument, and ground-truth measurements, in 1996 has been jointly supported by Matra Marconi Space France (MMS), the French Institut Technique des Céréales et des Fourrages (ITCF), and the French Institut National de la Recherche Agronomique (INRA).

References (58)

B Andrieu et al.
Evaluation of an improved version of SAIL model for simulating bi-directional reflectance of sugar beet canopies
Remote Sens. Environ.
(1997)
F Baret et al.
Potentials and limits of vegetation indices for LAI and APAR assessment
Remote Sens. Environ.
(1991)
F Baret et al.
Use of spectral analogy to evaluate canopy reflectance sensitivity to leaf optical properties
Remote Sens. Environ.
(1994)
A Bateson et al.
A method for manual endmember selection and spectral unmixing
Remote Sens. Environ.
(1996)
T.N Carlson et al.
On the relation between NDVI, fractional vegetation cover, and leaf area index
Remote Sens. Environ.
(1997)
E.W Chappelle et al.
Ratio analysis of reflectance spectra (RARS)an algorithm for remote estimation of the concentrations of chlorophyll A, chlorophyll B, and carotenoids in soybean leaves
Remote Sens. Environ.
(1992)
J.G.P.W Clevers
A simplified approach for yield prediction of sugar beet based on optical remote sensing data
Remote Sens. Environ.
(1997)
T Fourty et al.
Leaf optical properties with explicit description of its biochemical compositiondirect and inverse problems
Remote Sens. Environ.
(1996)
M.A Gilabert et al.
Analyses of spectral–biophysical relationships for a corn canopy
Remote Sens. Environ.
(1996)
L.D Hinzman et al.
Effects of nitrogen fertilization on growth and reflectance characteristics of winter wheat
Remote Sens. Environ.
(1986)

A.R Huete

Separation of soil-plant spectral mixtures by factor analysis

Remote Sens. Environ.

(1986)

A Huete et al.

Development of vegetation and soil indices for MODIS-EOS

Remote Sens. Environ.

(1994)

S.J Hurcom et al.

Assessment of biophysical vegetation properties through spectral decomposition techniques

Remote Sens. Environ.

(1996)

S Jacquemoud et al.

PROSPECTA model of leaf optical properties spectra

Remote Sens. Environ.

(1990)

S Jacquemoud

Inversion of the PROSPECT+SAIL canopy reflectance model from AVIRIS equivalent spectratheoretical study

Remote Sens. Environ.

(1993)

S Jacquemoud et al.

Estimating leaf biochemistry using the PROSPECT leaf optical properties model

Remote Sens. Environ.

(1996)

A Kuusk

A Markov chain canopy reflectance model

Agric. For. Meteorol.

(1995)

T.J Malthus et al.

Candidates high spectral resolution infrared indices for crop cover

Remote Sens. Environ.

(1993)

M.S Moran et al.

Opportunities and limitations for image-based remote sensing in precision crop management

Remote Sens. Environ.

(1997)

M Pax-Lenney et al.

The effect of spatial resolution on the ability to monitor the status of agricultural lands

Remote Sens. Environ.

(1997)

V.Y Railyan et al.

Red-edge structure of canopy reflectance spectra of triticale

Remote Sens. Environ.

(1993)

D.A Roberts et al.

Green vegetation, nonphotosynthetic vegetation, and soils in AVIRIS data

Remote Sens. Environ.

(1993)

D.A Roberts et al.

Temporal and spatial patterns in vegetation and atmospheric properties from AVIRIS

Remote Sens. Environ.

(1997)

B.F.T Rudorff et al.

Spectral response of wheat and its relationship to agronomic variables in the tropical region

Remote Sens. Environ.

(1990)

M.O Smith et al.

Vegetation in desertsI. A regional measure of abundance from multispectral images

Remote Sens. Environ.

(1990)

W Verhoef

Earth observation modeling based on layer scattering matrices

Remote Sens. Environ.

(1985)

M.M Verstraete et al.

Potential and limitations of information extraction on the terrestrial biosphere from satellite remote sensing

Remote Sens. Environ.

(1996)

C.L Wiegand et al.

Vegetation indices in crop assessments

Remote Sens. Environ.

(1991)

Adams, J. B., Smith, M. O., and Gillepsie, A. R. (1989), Simple models for complex natural surfaces: a strategy for the...

Cited by (158)

Mapping forest fire severity using bi-temporal unmixing of Sentinel-2 data - Towards a quantitative understanding of fire impacts
2023, Science of Remote Sensing
Precise quantification of forest fire impacts is critical for management strategies in support of post-fire mitigation. In this regard, optical remote sensing imagery in combination with spectral unmixing has been widely used to measure fire severity by means of fractional cover of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV), charcoal (CH) and further ground components such as ash, bare soil and rocks. However, most unmixing analyses have made use of a single post-fire image without accounting for the pre-fire state. We aim to assess fire severity from Sentinel-2 data using a bi-temporal spectral unmixing analysis that provides a quantitative fire impact description and is oriented towards the process of change by including pre-fire and post-fire information. Unmixing was based on Random Forest Regression (RFR) modeling using synthetic training data from a bi-temporal spectral library. We describe fire severity as changes associated with the combustion of photosynthetic vegetation (PV–CH fraction) and dieback of photosynthetic vegetation (PV-NPV fraction). Unburned forest was mapped as stable photosynthetic vegetation (PV-PV fraction). We evaluated our approach on a forest fire that burned in a temperate forest region in eastern Germany in 2018. Independent validation was carried out based on reference fractions obtained from very high-resolution (VHR) imagery such as Plante Scope, SPOT6, orthophotos, aerial photos, and Google Earth. The results underline the effectiveness of our unmixing approach, with Root Mean Squared Errors (RMSE) of 0.072 for PV-CH, 0.09 for PV-NPV, and 0.08 for PV-PV fractions. Most of the errors were caused by spectral similarity between charcoal and shadow effects caused by trees, and the coloring of foliage and NPV in the late phenological season of the post-fire Sentinel-2 image. Based on the two-dimensional feature space of PV-CH and PV-NPV fractions, we calculated two metrics to characterize fire impacts: distance, an indicator of disturbance severity (sum of combustion and dieback), and angle, a measure of disturbance composition (gradient between combustion and dieback). Furthermore, we compared the fraction-based metrics with the difference Normalized Burn Ratio (dNBR). Since the dNBR is most sensitive to combustion and presence of charcoal, it does not fully characterize fire-related vegetation loss associated with dieback. The bi-temporal fraction-based indices provide more ecologically meaningful information on fire severity, particularly for regions that are less prone to severe wildfires such as Central Europe.
The disposition of data from precision agricultural technologies: What do young agriculturalists think?
2023, Technology in Society
The absence of a legal and regulatory framework protecting a farmer's right to control data generated from their use of precision agriculture may impede adoption of the technology. Consideration of preferences about the disposition of this data is therefore critical to the broader discussion about the growth of precision agricultural technologies and its impact on society. Using a choice experiment, we show that factors contributing to willingness to enroll in a data management contract with a precision agriculture service provider include discount received, whether data ownership rights are retained, data privacy guarantees, and whether the data is transferred manually or automatically between systems.
A novel approach to combine spatial and spectral information from hyperspectral images
2023, Chemometrics and Intelligent Laboratory Systems
This article proposes a generic framework to process jointly the spatial and spectral information of hyperspectral images. First, sub-images are extracted. Then each of these sub-images follows two parallel workflows, one dedicated to the extraction of spatial features and the other dedicated to the extraction of spectral features. Finally, the extracted features are merged, producing as many scores as sub-images. Two applications are proposed, illustrating different spatial and spectral processing methods. The first one is related to the characterization of a teak wood disk, in an unsupervised way. It implements tensors of structure for the spatial branch, simple averaging for the spectral branch and multi-block principal component analysis for the fusion process. The second application is related to the early detection of apple scab on leaves. It implements co-occurrence matrices for the spatial branch, singular value decomposition for the spectral branch and multiblock partial least squares discriminant analysis for the fusion process. Both applications demonstrate the interest of the proposed method for the extraction of relevant spatial and spectral information and show how promising this new approach is for hyperspectral imaging processing.
Crop nitrogen monitoring: Recent progress and principal developments in the context of imaging spectroscopy missions
2020, Remote Sensing of Environment
Nitrogen (N) is considered as one of the most important plant macronutrients and proper management of N therefore is a pre-requisite for modern agriculture. Continuous satellite-based monitoring of this key plant trait would help to understand individual crop N use efficiency and thus would enable site-specific N management. Since hyperspectral imaging sensors could provide detailed measurements of spectral signatures corresponding to the optical activity of chemical constituents, they have a theoretical advantage over multi-spectral sensing for the detection of crop N. The current study aims to provide a state-of-the-art overview of crop N retrieval methods from hyperspectral data in the agricultural sector and in the context of future satellite imaging spectroscopy missions. Over 400 studies were reviewed for this purpose, identifying those estimating mass-based N (N concentration, N%) and area-based N (N content, N_area) using hyperspectral remote sensing data. Retrieval methods of the 125 studies selected in this review can be grouped into: (1) parametric regression methods, (2) linear nonparametric regression methods or chemometrics, (3) nonlinear nonparametric regression methods or machine learning regression algorithms, (4) physically-based or radiative transfer models (RTM), (5) use of alternative data sources (sun-induced fluorescence, SIF) and (6) hybrid or combined techniques. Whereas in the last decades methods for estimation of N_area and N% from hyperspectral data have been mainly based on simple parametric regression algorithms, such as narrowband vegetation indices, there is an increasing trend of using machine learning, RTM and hybrid techniques. Within plants, N is invested in proteins and chlorophylls stored in the leaf cells, with the proteins being the major nitrogen-containing biochemical constituent. However, in most studies, the relationship between N and chlorophyll content was used to estimate crop N, focusing on the visible-near infrared (VNIR) spectral domains, and thus neglecting protein-related N and reallocation of nitrogen to non-photosynthetic compartments. Therefore, we recommend exploiting the estimation of nitrogen via the proxy of proteins using hyperspectral data and in particular the short-wave infrared (SWIR) spectral domain. We further strongly encourage a standardization of nitrogen terminology, distinguishing between N% and N_area. Moreover, the exploitation of physically-based approaches is highly recommended combined with machine learning regression algorithms, which represents an interesting perspective for future research in view of new spaceborne imaging spectroscopy sensors.
Temperature determination in laser welding based upon a hyperspectral imaging technique
2019, CIRP Annals
Hyperspectral imaging (HSI) is a very promising technique for temperature determination in laser welding without the need for a priori knowledge of the emissivity. HSI is commonly considered a technique that allows to simultaneously acquire spatial as well as spectral information – that is essentially containing the temperature information – upon the investigated object. For that purpose, we developed a suitable HSI-system comprising a self-designed HSI-lens system and a high-speed CMOS-camera. In the present paper we address influences upon the HSI-based temperature determination and present as well as discuss temperature determination results in laser welding of metal sheets.
Big GIS analytics framework for agriculture supply chains: A literature review identifying the current trends and future perspectives
2018, Computers and Electronics in Agriculture
Citation Excerpt :
The hyperspectral and multispectral images obtained through the geospatial data is found to be very useful for analyzing parameters such as crop health and soil moisture. The GIS supports high level of decision making for effective management of fertilizer and pesticides, stress mapping, and irrigation (Barnes and Baker, 2000; Barroso et al., 2008; Hinzman et al., 1986; Lelong et al., 1998; Pal and Mather, 2003; Singh et al., 2007; Tilling et al., 2007). Bio-physical attributes of crops and soils are used with GIS and RS technologies for radical improvement in the agricultural productivity (Liaghat and Balasundram, 2010).
The world population is estimated to reach nine billion by 2050. Many challenges are adding pressure on the current agriculture supply chains that include shrinking land sizes, ever increasing demand for natural resources and environmental issues. The agriculture systems need a major transformation from the traditional practices to precision agriculture or smart farming practices to overcome these challenges. Geographic information system (GIS) is one such technology that pushes the current methods to precision agriculture. In this paper, we present a systematic literature review (SLR) of 120 research papers on various applications of big GIS analytics (BGA) in agriculture. The selected papers are classified into two broad categories; the level of analytics and GIS applications in agriculture. The GIS applications viz., land suitability, site search and selection, resource allocation, impact assessment, land allocation, and knowledge-based systems are considered in this study. The outcome of this study is a proposed BGA framework for agriculture supply chain. This framework identifies big data analytics to play a significant role in improving the quality of GIS application in agriculture and provides the researchers, practitioners, and policymakers with guidelines on the successful management of big GIS data for improved agricultural productivity.

View all citing articles on Scopus

View full text