Modeling and Prediction of Time-Varying Environmental Data Using Advanced Bayesian Methods

in Exploring Innovative and Successful Applications of Soft Computing (2013)

The problem of state/parameter estimation represents a key issue in crop models which are nonlinear, non-Gaussian and include a large number of parameters. The prediction errors are often important due to ... [more ▼]

The problem of state/parameter estimation represents a key issue in crop models which are nonlinear, non-Gaussian and include a large number of parameters. The prediction errors are often important due to uncertainties in the equations, the input variables, and the parameters. The measurements needed to run the model (input data), to perform calibration and validation are sometimes not numerous or known with some uncertainty. In these cases, estimating the state variables and/or parameters from easily obtained measurements can be extremely useful. In this work, we address the problem of modeling and prediction of leaf area index and soil moisture (LSM) using state estimation. The performances of various conventional and state-of-the-art state estimation techniques are compared when they are utilized to achieve this objective. These techniques include the extended Kalman filter (EKF), unscented Kalman filter (UKF), particle filter (PF), and the more recently developed technique variational Bayesian filter (VF). The objective of this work is to extend the state and parameter estimation techniques (i.e., EKF, UKF, PF and VF) to better handle nonlinear and non-Gaussian processes without a priori state information, by utilizing a time-varying assumption of statistical parameters. In this case, the state vector to be estimated at any instant is assumed to follow a Gaussian model, where the expectation and the covariance matrix are both random. The randomness of the expectation and the covariance of the state/parameter vector are assumed here to further capture the uncertainty of the state distribution. One practical choice of these distributions can be a Gaussian distribution for the expectation and a multi-dimensional Wishart distribution for the covariance matrix. The assumption of random mean and random covariance of the state leads to a probability distribution covering a wide range of tail behaviors, which allows discrete jumps in the state variables. [less ▲]

Vers un système de prédiction du rendement en temps réel

Poster (2012, October 16)

L'objectif de cette recherche est le développement d'un outil capable de prédire les rendements d'une culture de blé en temps réel, au fur et à mesure que la saison avance. Pour atteindre cet objectif ... [more ▼]

L'objectif de cette recherche est le développement d'un outil capable de prédire les rendements d'une culture de blé en temps réel, au fur et à mesure que la saison avance. Pour atteindre cet objectif, nous avons développé une méthodologie qui repose sur l'adjonction des éléments suivants : (i) le modèle de culture STICS (INRA, France), (ii) un réseau de capteurs sans fil pour le monitoring des variables agro-environnementales (éKo system, The Crossbow technologies, USA) et (iii) une base de données météorologiques. [less ▲]

Comparison of the effect of the thermal radiation from the sky or from Ecotron walls on the crop temperature

Conference (2012, June 01)

A first step towards a real-time predictive yield support system.

Conference (2012)

Real-time non-invasive determination of crop biomass and yield prediction are maybe among the major challenges in agriculture. But unknown future weather remains the key point of accurate yield forecast ... [more ▼]

Real-time non-invasive determination of crop biomass and yield prediction are maybe among the major challenges in agriculture. But unknown future weather remains the key point of accurate yield forecast. This paper presents the results of a preliminary study that aims to supply the unknown future by daily mean climatic data. The results show that, under the Belgian weather, this approach is relevant. Furthermore, the developed methodology appears to be a powerful diagnosis tool of the remaining yield potential under ongoing weather. [less ▲]

Bayesian methods for predicting LAI and soil moisture

in Proceedings of the 11th International Conference on Precision Agriculture (2012)

The prediction errors of crop models are often important due to uncertainties in the estimates of initial values of the states, in parameters, and in equations. The measurements needed to run the model ... [more ▼]

The prediction errors of crop models are often important due to uncertainties in the estimates of initial values of the states, in parameters, and in equations. The measurements needed to run the model are sometimes not numerous or known with some uncertainty. In this paper, two Bayesian filtering methods were used to update the state variable values predicted by MiniSTICS model. The chosen state variates were the LAI (Leaf Area Index) of a wheat crop (Triticum aestivum L.) and the corresponding moisture content of two soil layers (0-20 cm and 30-50 cm). These state variates were estimated simultaneously with several parameters. The assessed filtering methods were the centralized Particle Filtering (PF) and the Variational Bayesian Filtering (VF). The former is known to be sensitive to the number of particles while the latter yields an optimal choice of the sampling distribution over the state variable by minimizing the Kullback-Leibler divergence. In fact, variational calculus leads to a simple Gaussian sampling distribution whose parameters (estimated iteratively) depends on the observed data. On basis of a case study, the VF method was found more efficient than the PF method. Indeed, with the VF, the Root Mean Square Error (RMSE) of the three estimated states was smaller and the convergence of the all parameters was ensured. [less ▲]

La vision artificielle: une méthode d'avenir pour la reconnaissance automatisée des plantes adventices?

in Biotechnologie, Agronomie, Société et Environnement = Biotechnology, Agronomy, Society and Environment [=BASE] (2011), 15(4), 597-610

Weeds compete with crop plants for sunlight, moisture and nutrients and can have a detrimental impact on crop yields and quality if uncontrolled. They are destroyed by chemical, non chemical and ... [more ▼]

Weeds compete with crop plants for sunlight, moisture and nutrients and can have a detrimental impact on crop yields and quality if uncontrolled. They are destroyed by chemical, non chemical and integrated methods. To perform a site-specific weeds destruction, combination of these techniques with ground-based machine vision technology has high potential. Several methods exist to differentiate weeds from soil, between the rows. The more complicated problem is encountered when weeds are mixed to crops within the rows. Algorithms based on colorimetric or shape features are widely dependant on the variability of weeds and crops and are difficult to transpose from one situation to another. Measurement of plant height is a promising method, since at low spatial scale, the growthing speed is more uniform for the plants than for the weeds. This growing speed is function of the height and of a characteristic time, such as the number of days after sowing. To implement this method, active stereoscopy combined to an accurate measurement of the soil microrelief is required. [less ▲]

A MODEL FOR WHEAT YIELD PREDICTION BASED ON REAL-TIME MONITORING OF ENVIRONMENTAL FACTORS

Poster (2010, July)

This paper presents the results of a one year preliminary study in which a real time monitoring system was used to feed the STICS soil crop model. As indicated by the statistical criteria (RMSE and model ... [more ▼]

This paper presents the results of a one year preliminary study in which a real time monitoring system was used to feed the STICS soil crop model. As indicated by the statistical criteria (RMSE and model efficiency), the optimization of some wheat crop parameters allows the model to predict the yields with good accuracy for different soil type and different nitrogen application rates. [less ▲]