Inter-comparison of hemispherical conical reflectance factors (HCRF) measured with four fibre-based spectrometers

Unmanned Aerial Systems (UAS)-Based Methods for Solar Induced Chlorophyll Fluorescence (SIF) Retrieval with Non-Imaging Spectrometers: State of the Art

Chlorophyll fluorescence (ChlF) information offers a deep insight into the plant physiological status by reason of the close relationship it has with the photosynthetic activity. The unmanned aerial systems (UAS)-based assessment of solar induced ChlF (SIF) using non-imaging spectrometers and radiance-based retrieval methods, has the potential to provide spatio-temporal photosynthetic performance information at field scale. The objective of this manuscript is to report the main advances in the development of UAS-based methods for SIF retrieval with non-imaging spectrometers through the latest scientific contributions, some of which are being developed within the frame of the Training on Remote Sensing for Ecosystem Modelling (TRuStEE) program. Investigations from the Universities of Edinburgh (School of Geosciences) and Tasmania (School of Technology, Environments and Design) are first presented, both sharing the principle of the spectroradiometer optical path bifurcation throughout, the so called ‘Piccolo-Doppio’ and ‘AirSIF’ systems, respectively. Furthermore, JB Hyperspectral Devices’ ongoing investigations towards the closest possible characterization of the atmospheric interference suffered by orbital platforms are outlined. The latest approach focuses on the observation of one single ground point across a multiple-kilometer atmosphere vertical column using the high altitude UAS named as AirFloX, mounted on a specifically designed and manufactured fixed wing platform: ‘FloXPlane’. We present technical details and preliminary results obtained from each instrument, a summary of their main characteristics, and finally the remaining challenges and open research questions are addressed. On the basis of the presented findings, the consensus is that SIF can be retrieved from low altitude spectroscopy. However, the UAS-based methods for SIF retrieval still present uncertainties associated with the current sensor characteristics and the spatio-temporal mismatching between aerial and ground measurements, which complicate robust validations. Complementary studies regarding the standardization of calibration methods and the characterization of spectroradiometers and data processing workflows are also required. Moreover, other open research questions such as those related to the implementation of atmospheric correction, bidirectional reflectance distribution function (BRDF) correction, and accurate surface elevation models remain to be addressed.

A Machine Learning Framework to Predict Nutrient Content in Valencia-Orange Leaf Hyperspectral Measurements

This paper presents a framework based on machine learning algorithms to predict nutrient content in leaf hyperspectral measurements. This is the first approach to evaluate macro- and micronutrient content with both machine learning and reflectance/first-derivative data. For this, citrus-leaves collected at a Valencia-orange orchard were used. Their spectral data was measured with a Fieldspec ASD FieldSpec® HandHeld 2 spectroradiometer and the surface reflectance and first-derivative spectra from the spectral range of 380 to 1020 nm (640 spectral bands) was evaluated. A total of 320 spectral signatures were collected, and the leaf-nutrient content (N, P, K, Mg, S, Cu, Fe, Mn, and Zn) was associated with them. For this, 204,800 (320 × 640) combinations were used. The following machine learning algorithms were used in this framework: k-Nearest Neighbor (kNN), Lasso Regression, Ridge Regression, Support Vector Machine (SVM), Artificial Neural Network (ANN), Decision Tree (DT), and Random Forest (RF). The training methods were assessed based on Cross-Validation and Leave-One-Out. The Relief-F metric of the algorithms’ prediction was used to determine the most contributive wavelength or spectral region associated with each nutrient. This approach was able to return, with high predictions (R2), nutrients like N (0.912), Mg (0.832), Cu (0.861), Mn (0.898), and Zn (0.855), and, to a lesser extent, P (0.771), K (0.763), and S (0.727). These accuracies were obtained with different algorithms, but RF was the most suitable to model most of them. The results indicate that, for the Valencia-orange leaves, surface reflectance data is more suitable to predict macronutrients, while first-derivative spectra is better linked to micronutrients. A final contribution of this study is the identification of the wavelengths responsible for contributing to these predictions.

Modeling Hyperspectral Response of Water-Stress Induced Lettuce Plants Using Artificial Neural Networks

Modeling the hyperspectral response of vegetables is important for estimating water stress through a noninvasive approach. This article evaluates the hyperspectral response of water-stress induced lettuce (Lactuca sativa L.) using artificial neural networks (ANN). We evenly split 36 lettuce pots into three groups: control, stress, and bacteria. Hyperspectral response was measured four times, during 14 days of stress induction, with an ASD Fieldspec HandHeld spectroradiometer (325–1075 nm). Both reflectance and absorbance measurements were calculated. Different biophysical parameters were also evaluated. The performance of the ANN approach was compared against other machine learning algorithms. Our results show that the ANN approach could separate the water-stressed lettuce from the non-stressed group with up to 80% accuracy at the beginning of the experiment. Additionally, this accuracy improved at the end of the experiment, reaching an accuracy of up to 93%. Absorbance data offered better accuracy than reflectance data to model it. This study demonstrated that it is possible to detect early stages of water stress in lettuce plants with high accuracy based on an ANN approach applied to hyperspectral data. The methodology has the potential to be applied to other species and cultivars in agricultural fields.

Sun-Induced Chlorophyll Fluorescence I: Instrumental Considerations for Proximal Spectroradiometers

Growing interest in the proximal sensing of sun‐induced chlorophyll fluorescence (SIF) has been boosted by space-based retrievals and up-coming missions such as the FLuorescence EXplorer (FLEX). The European COST Action ES1309 “Innovative optical tools for proximal sensing of ecophysiological processes” (OPTIMISE, ES1309; https://optimise.dcs.aber.ac.uk/) has produced three manuscripts addressing the main current challenges in this field. This article provides a framework to model the impact of different instrument noise and bias on the retrieval of SIF; and to assess uncertainty requirements for the calibration and characterization of state-of-the-art SIF-oriented spectroradiometers. We developed a sensor simulator capable of reproducing biases and noises usually found in field spectroradiometers. First the sensor simulator was calibrated and characterized using synthetic datasets of known uncertainties defined from laboratory measurements and literature. Secondly, we used the sensor simulator and the characterized sensor models to simulate the acquisition of atmospheric and vegetation radiances from a synthetic dataset. Each of the sensor models predicted biases with propagated uncertainties that modified the simulated measurements as a function of different factors. Finally, the impact of each sensor model on SIF retrieval was analyzed. Results show that SIF retrieval can be significantly affected in situations where reflectance factors are barely modified. SIF errors were found to correlate with drivers of instrumental-induced biases which are as also drivers of plant physiology. This jeopardizes not only the retrieval of SIF, but also the understanding of its relationship with vegetation function, the study of diel and seasonal cycles and the validation of remote sensing SIF products. Further work is needed to determine the optimal requirements in terms of sensor design, characterization and signal correction for SIF retrieval by proximal sensing. In addition, evaluation/validation methods to characterize and correct instrumental responses should be developed and used to test sensors performance in operational conditions.