Hyperspectral images of grapevine leaves including healthy leaves and leaves with biotic and abiotic symptoms.

A hyperspectral imaging database was collected on two hundred and five grape plant leaves. Leaves were measured with a hyperspectral camera in the visible/near infrared spectral range under controlled conditions. This dataset contains hyperspectral acquisition of grape leaves of seven different varieties. For each variety, acquisitions were performed on healthy leaves and leaves with foliar symptoms caused by different grapevine diseases showing clear symptoms of biotic or abiotic stress on other organs. For each leaf, chemical measurements such as chlorophyll and flavonol contents were also performed.

Dataset containing spectral data from hyperspectral imaging and sugar content measurements of grapes berries in various maturity stage.

In the dataset presented in this article, two hundred and seventy four trays containing one hundred berries were measured by a hyperspectral camera in the visible/near-infrared spectral domain. This dataset was formed to study the use of hyperspectral imaging for maturity monitoring of grape berries [2]. This dataset contains reflectance spectra from hyperspectral camera of grape berries of three different varieties and chemical composition (sugar content).

Discriminating between Absorption and Scattering Effects in Complex Turbid Media by Coupling Polarized Light Spectroscopy with the Mueller Matrix Concept.

The separation of the combined effects of absorption and scattering in complex media is a major issue for better characterization and prediction of media properties. In this study, an approach coupling polarized light spectroscopy and the Mueller matrix concept were evaluated to address this issue. A set of 50 turbid liquid optical phantoms with different levels of scattering and absorption properties were made and measured at various orientations of polarizers and analyzers to obtain the 16 elements of the complete Mueller matrix in the VIS-NIR region. Partial least square (PLS) was performed to build calibration models from diffuse reflectance spectra in order to evaluate the potential of polarization spectroscopy through the elements of the Mueller matrix to predict physical and chemical parameters and hence, to discriminate scattering and absorption effects, respectively. In particular, it was demonstrated that absorption and scattering effects can be distinguished in the Rayleigh regime with linear and circular polarization from the M22 and M44 elements of the Mueller matrix, correspondingly.

Terahertz probing of sunflower leaf multilayer organization.

We analyze the multilayer structure of sunflower leaves from Terahertz data measured in the time-domain at a ps scale. Thin film reverse engineering techniques are applied to the Fourier amplitude of the reflected and transmitted signals in the frequency range f < 1.5 Terahertz (THz). Validation is first performed with success on etalon samples. The optimal structure of the leaf is found to be a 8-layer stack, in good agreement with microscopy investigations. Results may open the door to a complementary classification of leaves.

A New Optical Sensor Based on Laser Speckle and Chemometrics for Precision Agriculture: Application to Sunflower Plant-Breeding.

New instruments to characterize vegetation must meet cost constraints while providing accurate information. In this paper, we study the potential of a laser speckle system as a low-cost solution for non-destructive phenotyping. The objective is to assess an original approach combining laser speckle with chemometrics to describe scattering and absorption properties of sunflower leaves, related to their chemical composition or internal structure. A laser diode system at two wavelengths 660 nm and 785 nm combined with polarization has been set up to differentiate four sunflower genotypes. REP-ASCA was used as a method to analyze parameters extracted from speckle patterns by reducing sources of measurement error. First findings have shown that measurement errors are mostly due to unwilling residual specular reflections. Moreover, results outlined that the genotype significantly impacts measurements. The variables involved in genotype dissociation are mainly related to scattering properties within the leaf. Moreover, an example of genotype classification using REP-ASCA outcomes is given and classify genotypes with an average error of about 20%. These encouraging results indicate that a laser speckle system is a promising tool to compare sunflower genotypes. Furthermore, an autonomous low-cost sensor based on this approach could be used directly in the field.

Dataset of visible-near infrared handheld and micro-spectrometers - comparison of the prediction accuracy of sugarcane properties.

In the dataset presented in this article, sixty sugarcane samples were analyzed by eight visible / near infrared spectrometers including seven micro-spectrometers. There is one file per spectrometer with sample name, wavelength, absorbance data [calculated as log10 (1/Reflectance)], and another file for reference data, in order to assess the potential of the micro-spectrometers to predict chemical properties of sugarcane samples and to compare their performance with a LabSpec spectrometer. The Partial Least Square Regression (PLS-R) algorithm was used to build calibration models. This open access dataset could also be used to test new chemometric methods, for training, etc.

Combining light polarization and speckle measurements with multivariate analysis to predict bulk optical properties of turbid media.

This study aims to investigate the combination of speckle pattern analysis, polarization parameters, and chemometric tools to predict the optical absorption and scattering properties of materials. For this purpose, an optical setup based on light polarization and speckle measurements was developed, and turbid samples were measured at 405 and 660 nm. First, a backscattered polarized speckle acquisition was performed on a set of 41 samples with various scattering (${\mu}_s$µs) and absorbing (${{\mu}_a}$µa) coefficients. Then, several parameters were computed from the polarized speckle images, and prediction models were built using stepwise multiple linear regression. For scattering media, ${{\mu}_s}$µs was predicted with ${R^{2} = 0.9}$R2=0.9 using two parameters. In the case of scattering and absorbing media, prediction results using two parameters were ${R^{2} = 0.62}$R2=0.62 for ${{\mu}_s}$µs and ${R^{2} = 0.8}$R2=0.8 for ${{\mu}_a}$µa. The overall results obtained in this research showed that the combination of speckle pattern analysis, polarization parameters, and chemometric tools to predict the optical bulk properties of materials show interesting promise.

Coupling Waveguide-Based Micro-Sensors and Spectral Multivariate Analysis to Improve Spray Deposit Characterization in Agriculture.

The leaf coverage surface is a key measurement of the spraying process to maximize spray efficiency. To determine leaf coverage surface, the development of optical micro-sensors that, coupled with a multivariate spectral analysis, will be able to measure the volume of the droplets deposited on their surface is proposed. Rib optical waveguides based on Ge-Se-Te chalcogenide films were manufactured and their light transmission was studied as a response to the deposition of demineralized water droplets on their surface. The measurements were performed using a dedicated spectrophotometric bench to record the transmission spectra at the output of the waveguides, before (reference) and after drop deposition, in the wavelength range between 1200 and 2000 nm. The presence of a hollow at 1450 nm in the relative transmission spectra has been recorded. This corresponds to the first overtone of the O-H stretching vibration in water. This result tends to show that the optical intensity decrease observed after droplet deposition is partly due to absorption by water of the light energy carried by the guided mode evanescent field. The probe based on Ge-Se-Te rib optical waveguides is thus sensitive throughout the whole range of volumes studied, i.e., from 0.1 to 2.5 µL. Principal Component Analysis and Partial Least Square as multivariate techniques then allowed the analysis of the statistics of the measurements and the predictive character of the transmission spectra. It confirmed the sensitivity of the measurement system to the water absorption, and the predictive model allowed the prediction of droplet volumes on an independent set of measurements, with a correlation of 66.5% and a precision of 0.39 µL.