Search results for "Mean squared error"

Fusing optical and SAR time series for LAI gap filling with multioutput Gaussian processes

2019

The availability of satellite optical information is often hampered by the natural presence of clouds, which can be problematic for many applications. Persistent clouds over agricultural fields can mask key stages of crop growth, leading to unreliable yield predictions. Synthetic Aperture Radar (SAR) provides all-weather imagery which can potentially overcome this limitation, but given its high and distinct sensitivity to different surface properties, the fusion of SAR and optical data still remains an open challenge. In this work, we propose the use of Multi-Output Gaussian Process (MOGP) regression, a machine learning technique that learns automatically the statistical relationships among…

Comparing different profiles to characterize the atmosphere for three MODIS TIR bands

2015

Abstract Accurate land surface temperature (LST) retrievals from sensors aboard orbiting satellites are dependent on the corresponding atmospheric correction, especially in the thermal infrared (TIR) spectral domain (8–14 μm). To remove the atmospheric effects from at-sensor measured radiance in the TIR range it is needed to characterize the atmosphere by means of three specific variables: the upwelling path and the hemispherical downwelling radiances plus the atmospheric transmissivity. Those variables can be derived from the previous knowledge of vertical atmospheric profiles of air temperature and relative humidity at different geo-potential heights and pressures. In this work, the above…

Application of the group method of data handling (GMDH) approach for landslide susceptibility zonation using readily available spatial covariates

2022

Abstract Landslide susceptibility (LS) mapping is an essential tool for landslide risk assessment. This study aimed to provide a new approach with better performance for landslide mapping and adopting readily available variables. In addition, it investigates the capability of a state-of-the-art model developed using the group method of data handling (GMDH) to spatially model LS. Furthermore, hybridized models of GMDH were developed using different metaheuristic algorithms. The study area was the Bonghwa region of South Korea, for which an accurate landslide inventory dataset is available. We considered a total of 13 spatial covariates (altitude, slope, aspect, topographic wetness index, val…

Prediction of organic carbon and total nitrogen contents in organic wastes and their composts by Infrared spectroscopy and partial least square regre…

2017

Middle and near infrared (MIR and NIR) were employed to determine organic carbon (OC) and total nitrogen (TN) in different soil organic amendments including wastes, composts and mixtures of composts and organic wastes. Prediction models based on partial least squares (PLS) regression from the spectra of untreated samples were built. Different spectra preprocessing strategies were adopted and the best number of latent variable was evaluated using leave-one-out cross-validation. Attenuated total reflectance (PLS-ATR-MIR) and diffuse reflectance (PLS-DR-NIR) models were built and evaluated from root mean square error of cross validation and prediction (RMSECV and RMSEP), coefficients of determ…

2004

This paper presents the use of Support Vector Machines (SVMs) for prediction and analysis of antisense oligonucleotide (AO) efficacy. The collected database comprises 315 AO molecules including 68 features each, inducing a problem well-suited to SVMs. The task of feature selection is crucial given the presence of noisy or redundant features, and the well-known problem of the curse of dimensionality. We propose a two-stage strategy to develop an optimal model: (1) feature selection using correlation analysis, mutual information, and SVM-based recursive feature elimination (SVM-RFE), and (2) AO prediction using standard and profiled SVM formulations. A profiled SVM gives different weights to …

Intelligent Sampling for Vegetation Nitrogen Mapping Based on Hybrid Machine Learning Algorithms

2021

Upcoming satellite imaging spectroscopy missions will deliver spatiotemporal explicit data streams to be exploited for mapping vegetation properties, such as nitrogen (N) content. Within retrieval workflows for real-time mapping over agricultural regions, such crop-specific information products need to be derived precisely and rapidly. To allow fast processing, intelligent sampling schemes for training databases should be incorporated to establish efficient machine learning (ML) models. In this study, we implemented active learning (AL) heuristics using kernel ridge regression (KRR) to minimize and optimize a training database for variational heteroscedastic Gaussian processes regression (V…

Retrieval of vegetation height in rice fields using polarimetric SAR interferometry with TanDEM-X data

2017

This work presents for the first time a demonstration with satellite data of polarimetric SAR interferometry (PolInSAR) applied to the retrieval of vegetation height in rice fields. Three series of dual-pol interferometric SAR data acquired with large baselines (2–3 km) by the TanDEM-X system during its science phase (April–September 2015) are exploited. A novel inversion algorithm especially suited for rice fields cultivated in flooded soil is proposed and evaluated. The validation is carried out over three test sites located in geographically different areas: Sevilla (SW Spain), Valencia (E Spain), and Ipsala (W Turkey), in which different rice types are present. Results are obtained duri…

A priori parameterisation of the CERES soil-crop models and tests against several European data sets

2002

Mechanistic soil-crop models have become indispensable tools to investigate the effect of management practices on the productivity or environmental impacts of arable crops. Ideally these models may claim to be universally applicable because they simulate the major processes governing the fate of inputs such as fertiliser nitrogen or pesticides. However, because they deal with complex systems and uncertain phenomena, site-specific calibration is usually a prerequisite to ensure their predictions are realistic. This statement implies that some experimental knowledge on the system to be simulated should be available prior to any modelling attempt, and raises a tremendous limitation to practica…

Performances of neural networks for deriving LAI estimates from existing CYCLOPES and MODIS products

2008

International audience; This paper evaluates the performances of a neural network approach to estimate LAI from CYCLOPES and MODIS nadir normalized reflectance and LAI products. A data base was generated from these products over the BELMANIP sites during the 2001-2003 period. Data were aggregated at 3 km x 3 km, resampled at 1/16 days temporal frequency and filtered to reject outliers. VEGETATION and MODIS reflectances show very consistent values in the red, near infrared and short wave infrared bands. Neural networks were trained over part of this data base for each of the 6 MODIS biome classes to retrieve both MODIS and CYCLOPES LAI products. Results show very good performances of neural …

Field testing of repurposed electric vehicle batteries for price-driven grid balancing

2019

Abstract As electric cars become more widespread, the disposal and recycling of used batteries will become an important challenge. Typically, vehicle batteries are replaced if their capacity drops to 70–80% of initial capacity. However, they may still be useful for stationary applications. In this paper, results from a field test of a molten salt high-temperature electric vehicle battery repurposed as stationary storage for grid balancing are presented. In a previous study, we have shown that a mixed integer linear programming control strategy driven by a spot-market price for electricity is best suited for an implementation on hardware with limited computational resources. A 14-day experim…