Using a mobile real-time soil visible-near infrared sensor for high resolution soil property mapping

The application of visible and near‐infrared (vis–NIR) spectroscopy to characterize soil samples has gained growing interest as a fast and cost‐effective methodology for soil fertility assessment. In order to profit from the full potential of vis–NIR spectroscopy, the acquisition of soil spectra directly in‐situ would increase the possibility to obtain data rapidly and at a high spatial and temporal resolution. In the present study, we test and propose the best practice to characterize a set of fertility‐related parameters (i.e. texture, organic carbon, pH, cation exchange capacity and major nutrients) of agricultural soils by measuring vis–NIR spectra in the field. To reach this goal, we compare the spectra obtained from different scanning positions with two portable spectrometers, that is, a micro‐electro‐mechanical systems (MEMS)‐based spectrometer and a research‐grade vis–NIR spectrometer. On the basis of 134 soil sampling points, vis–NIR spectra were recorded from: (1) the cutaway side of a soil sample collected with an Edelman auger to a depth of 20 cm, (2) the raw soil surface, as well as (3) the cleaned and smoothed soil surface. Partial least squares regression (PLSR) calibration models were built for the selected soil parameters, scanning positions and different spectral pretreatments for both spectrometers. The model performance was evaluated based on the ratio of performance to interquartile range (RPIQ), the R 2 , the root mean squared error (RMSE) and Lin's concordance correlation coefficient (CCC). Overall, the following soil parameters were successfully predicted: clay, sand, pH, organic carbon, cation exchange capacity, total nitrogen and exchangeable magnesium. In contrast, total and exchangeable Ca, K and P, as well as total Mg could not be predicted at a satisfactory level for both the spectrometers. The best scanning position for the successfully calibrated models was along the cutaway sides of the Edelman auger. Although the research‐grade spectrometer gave better performance indicators for most of the parameters, the calibrations with the MEMS‐based spectrometer still resulted in satisfactory predictions. Based on these findings, the proposed best practice for obtaining in‐situ soil vis–NIR scans is to scan along the cutaway sides of a soil core using at least five replicate scans.

<p>The Joint IAEA/FAO Division of Nuclear Techniques in Food and Agriculture, through its Soil and Water Management & Crop Nutrition Laboratory (SWMCNL), launched in October 2019, a new Coordinated Research Project (D15019) called “Monitoring and Predicting Radionuclide Uptake and Dynamics for Optimizing Remediation of Radioactive Contamination in Agriculture''. Within this context, the high-throughput characterization of soil properties in general and the estimation of soil-to-plant transfer factors of radionuclides are of critical importance.</p><p>For several decades, soil researchers have been successfully using near and mid-infrared spectroscopy (MIRS) techniques to estimate a wide range of soil physical, chemical and biological properties such as carbon (C), Cation Exchange Capacities (CEC), among others. However, models developed were often limited in scope as only small and region-specific MIR spectra libraries of soils were accessible.</p><p>This situation of data scarcity is changing radically today with the availability of large and growing library of MIR-scanned soil samples maintained by the National Soil Survey Center (NSSC) Kellogg Soil Survey Laboratory (KSSL) from the United States Department of Agriculture (USDA-NRCS) and the Global Soil Laboratory Network (GLOSOLAN) initiative of the Food Agency Organization (FAO). As a result, the unprecedented volume of data now available allows soil science researchers to increasingly shift their focus from traditional modeling techniques such as PLSR (Partial Least Squares Regression) to classes of modeling approaches, such as Ensemble Learning or Deep Learning, that have proven to outperform PLSR on most soil properties prediction in a large data regime.</p><p>As part of our research, the opportunity to train higher capacity models on the KSSL large dataset (all soil taxonomic orders included ~ 50K samples) makes it possible to reach a quality of prediction for exchangeable potassium so far unsurpassed with a Residual Prediction Deviation (RPD) around 3. Potassium is known for its difficulty of being predicted but remains extremely important in the context of remediation of radioactive contamination after a nuclear accident. Potassium can help reduce the uptake of radiocaesium by crops, as it competes with radiocaesium in soil-to-plant transfer.</p><p>To ensure informed decision making, we also guarantee that (i) individual predictions uncertainty is estimated (using Monte Carlo Dropout) and (ii) individual predictions can be interpreted (i.e. how much specific MIRS wavenumber regions contribute to the prediction) using methods such as Shapley Additive exPlanations (SHAP) values.</p><p>SWMCNL is now a member of the GLOSOLAN network, which helps enhance the usability of MIRS for soil monitoring worldwide. SWMCNL is further developing training packages on the use of traditional and advanced mathematical techniques to process MIRS data for predicting soil properties. This training package has been tested in October 2020 with thirteen staff members of the FAO/IAEA Laboratories in Seibersdorf, Austria.</p>

As a laboratory proximal sensing technique, the capability of visible and near-infrared (Vis-NIR) diffused reflectance spectroscopy with partial least squares (PLS) regression to determine soil properties has previously been demonstrated. However, the evaluation of the soil phosphorus (P) content—a major nutrient constraint for crop production in the tropics—is still a challenging task. PLS regression with waveband selection can improve the predictive ability of a calibration model, and a genetic algorithm (GA) has been widely applied as a suitable method for selecting wavebands in laboratory calibrations. To develop a laboratory-based proximal sensing method, this study investigated the potential to use GA-PLS regression analyses to estimate oxalate-extractable P in upland and lowland soils from laboratory Vis-NIR reflectance data. In terms of predictive ability, GA-PLS regression was compared with iterative stepwise elimination PLS (ISE-PLS) regression and standard full-spectrum PLS (FS-PLS) regression using soil samples collected in 2015 and 2016 from the surface of upland and lowland rice fields in Madagascar (n = 103). Overall, the GA-PLS model using first derivative reflectance (FDR) had the best predictive accuracy (R2 = 0.796) with a good prediction ability (residual predictive deviation (RPD) = 2.211). Selected wavebands in the GA-PLS model did not perfectly match wavelengths of previously known absorption features of soil nutrients, but in most cases, the selected wavebands were within 20 nm of previously known wavelength regions. Bootstrap procedures (N = 10,000 times) using selected wavebands also confirmed the improvements in accuracy and robustness of the GA-PLS model compared to those of the ISE-PLS and FS-PLS models. These results suggest that soil oxalate-extractable P can be predicted from Vis-NIR spectroscopy and that GA-PLS regression has the advantage of tuning optimum bands for PLS regression, contributing to a better predictive ability.

IntroductionAn accurate and reliable detection of soil physicochemical attributes (SPAs) is a difficult and complicated issue in soil science. The SPA may be varied spatially and temporally with the complexity of nature. In the past, SPA detection has been obtained through routine soil chemical and physical laboratory analysis. However, these laboratory methods do not fulfill the rapid requirements. Accordingly, diffuse reflectance spectroscopy (DRS) can be used to nondestructively detect and characterize soil attributes with superior solution. In the present article, we report a study done through spectral curves in the visible (350–700 nm) and near-infrared (700–2500 nm) (VNIR) region of 74 soil specimens which were agglomerated by farming sectors of Phulambri Tehsil of the Aurangabad region of Maharashtra, India. The quantitative analysis of VNIR spectrum was done.ResultsThe spectra of agglomerated farming soils were acquired by the Analytical Spectral Device (ASD) Field spec 4 spectroradiometer. The soil spectra of the VNIR region were preprocessed to get pure spectra which were the input for regression modeling. The partial least squares regression (PLSR) model was computed to construct the calibration models, which were individually validated for the prediction of SPA from the soil spectrum. The computed model was based on a correlation study between reflected spectra and detected SPA. The detected SPAs were soil organic carbon (SOC), nitrogen (N), soil organic matter (SOM), pH values, electrical conductivity (EC), phosphorus (P), potassium (K), iron (Fe), sand, silt, and clay. The accuracy of the PLSR model-validated determinant (R2) values were SOC 0.89, N 0.68, SOM 0.93, pH values 0.82, EC 0.89, P 0.98, K 0.82, Fe 0.94, sand 0.98, silt 0.90, and clay 0.69 with root mean square error of prediction (RMSEP) 3.51, 4.34, 2.66, 2.12, 4.11, 1.41, 4.22, 1.56, 1.89, 1.97, and 9.91, respectively. According to the experimental results, the VNIR-DRS was better for detection of SPA and produced more accurate predictions for SPA.ConclusionsIn conclusion, the methods examined here offered rapid and novel detection of SPA from reflectance spectroscopy. The outcome of the present research will be apt for precision farming and decision-making.

New approach for rapid estimation of leaf nitrogen, phosphorus, and potassium contents in apple-trees using Vis/NIR spectroscopy based on wavelength selection coupled with machine learning

Field experiment was conducted with the objective to determine the effect of inorganic fertilizer on soil properties, growth and yield of maize at the Teaching and Research Farm, Federal University of Agriculture, Makurdi, Southern Guinea Savanna Zone of Nigeria, during 2017 and 2018 cropping seasons. The experimental treatments consisted of 0, 75, 150 and 300 kg ha-1 of NPK fertilizer laid out in a randomized complete block design (RCBD) and replicated thrice. Soil samples were collected from a plough layer of 0-15 cm before and after the experiment for each cropping season according to the treatments and were analyzed for particle size distribution (PSD), soil pH, organic carbon (OC), total nitrogen (TN), available phosphorus, exchangeable cations and cation exchange capacity (CEC). Maize plant (growth and yield) data collected were subjected to analysis of variance after which significant means were separated using least significant difference (LSD) at P < 0.05. From the results, the effects of fertilizer treatments on soil properties investigated were significant. Application of fertilizer increased soil organic matter (SOM), TN and CEC in both cropping seasons. However, the SOM was low in all plots with no application of fertilizer. Exchangeable bases, base saturation and CEC reduced in 2017 in no fertilizer treatments but remarkably increased in 2018 due to application of 75, 150 and 300 kg ha-1. The 300 kgha-1 NPK fertilizer applications produced higher growth and yields of maize plant than remaining treatments. Similarly, soil properties (soil pH, SOM, TN and CEC) were also improved with application higher levels of mineral fertilizers for agricultural production.

با برآورد ظرفیت تبادل کاتیونی با استفاده از ویژگی‏های پایه ای و زودیافت خاک می‏توان در وقت و هزینه صرفه‏جویی کرد. هدف این بررسی، مقایسه دو روش رگرسیون درختی و رگرسیون خطی چندگانه در برآورد ظرفیت تبادل کاتیونی (CEC) با استفاده از ویژگی‏هایخاک است. برای این منظور از داده های 106 نمونه خاک UNSODA استفاده شد. جهت برآورد CEC با استفاده از روش‏های رگرسیون درختی و رگرسیون خطی از ویژگی های اجزای بافت خاک،pH، ماده آلی و چگالی ظاهری استفاده شد. کارآیی روش رگرسیون درختی در برابر روش رگرسیون خطی چندگانه در برآورد CEC مقایسه شدند. نتایج نشان داد که در برآورد CEC با استفاده از روش های رگرسیونی تنها ضرایب ماده آلی (183/3) و درصد رس (274/0) که بیشترین همبستگی را با CEC دارند معنی دار شده و مدل رگرسیونی بر اساس این دو پارامتر توسعه یافت. همچنین از میان پارمترهای ورودی در روش رگرسیون درختی تنها پارامترهای ماده آلی و درصد رس در درخت رگرسیون ظاهر شد. روش رگرسیون درختی در دو مجموعه داده صحت سنجی و اعتبار سنجی بر اساس آماره های ارزیابی R2، RMSE، ME و GMER کارآیی بالاتری نسبت به روش های رگرسیونی خطی در برآورد CEC داشت. از میان روش های رگرسیونی خطی، مدل پیشنهادی کارآیی بالاتری نسبت به مدل های بل و ونکولن و بروسما و همکاران داشت.

Zinc (Zn) desorption is an important process to determine Zn bioavailability in calcareous soils. An experiment was performed to assess the pattern of Zn release from 10 calcareous soils of orange orchards, southern Iran and the soil properties influencing it. For Zn desorption studies, soil samples were extracted with diethylene triamine penta-acetic acid solution at pH 7.3 for periods of 0.083–48 h. Suitability of seven kinetic models was also investigated to describe Zn release from soils. Generally, Zn desorption pattern was characterized by a rapid initial desorption up to 2 h of equilibration, followed by a slower release rate. The simple Elovich and two-constant rate kinetic models described Zn release the best, so it seems that Zn desorption is probably controlled by diffusion phenomena. The values of the rate constants for the superior models were significantly correlated with some soil properties such as soil organic matter (SOM) content, cation exchange capacity (CEC), and soil pH, whereas carbonate calcium equivalent and clay content had no significant influence on Zn desorption from soils. SOM had a positive effect on the magnitude of Zn release from soils, while soil pH showed a negative effect on Zn desorption. Furthermore, the initial release rate of soil Zn is probably controlled by CEC in the studied soils. Finally, it could be concluded that SOM, CEC, and soil pH are the most important factors controlling Zn desorption from calcareous soils of orange orchards, southern Iran. Abbreviations: Soil organic matter (SOM); Cation exchange capacity (CEC); Calcium carbonate equivalent (CCE); Zinc (Zn).

Choosing appropriate multivariate calibration and preprocessing transformation techniques is important in the determination of soil organic carbon (SOC) content based on visible and near-infrared (Vis–NIR) spectroscopy. The performance levels of partial least-squares regression (PLSR), support vector machine regression (SVMR), and wavelet neural network (WNN) calibration methods coupled with different preprocessing approaches were compared using three kinds of criteria, including the coefficient of determination (R2), root mean square error (RMSE), and residual prediction deviation (RPD). A total of 328 soil samples collected from the south bank of Hangzhou Bay were used as the dataset for the calibration–validation procedure and SOC content inversion. The effects of spectra preprocessing transformation methods were evaluated for raw spectra, Savitzky–Golay smoothing with the first derivatives of reflectance (FDR) and Savitzky–Golay smoothing with logarithm of reciprocal of the reflectance (log R−1). The results indicate that the SVMR is superior to the PLSR, and WNN models for SOC content prediction. The combination of the SVMR model with FDR provided the best prediction results for the SOC content, with R2p = 0.92, RPDP = 2.82, RMSEP = 0.36%, and a kappa correlation coefficient of interpolation as high as 0.97. The FDR of Vis–NIR spectroscopy combined with the SVMR model is recommended over the PLSR and WNN modeling techniques for the high-accuracy determination of the SOC content.

Rapid development in the glutamate fermentation industry has dictated the need for effective fermentation monitoring by rapid and precise methods that provide real-time information for quality control of the end-product. In recent years, near-infrared (NIR) spectroscopy and multivariate calibration have been developed as fast, inexpensive, non-destructive and environmentally safe techniques for industrial applications. The purpose of this study was to develop models for monitoring glutamate, glucose, lactate and alanine concentrations in the temperature-triggered process of glutamate fermentation. NIR measurements of eight batches of samples were analyzed by partial least-squares regression with several spectral pre-processing methods. The coefficient of determination (R (2)), model root-mean square error of calibration (RMSEC), root-mean square error of prediction (RMSEP) and residual predictive deviation (RPD) of the test calibration for the glutamate concentration were 0.997, 3.11 g/L, 2.56 g/L and 19.81, respectively. For the glucose concentration, R (2), RMSEC, RMSEP and RPD were 0.989, 1.37 g/L, 1.29 g/L and 9.72, respectively. For the lactate concentration, R (2), RMSEC, RMSEP and RPD were 0.975, 0.078 g/L, 0.062 g/L and 6.29, respectively. For the alanine concentration, R (2), RMSEC, RMSEP and RPD were 0.964, 0.213 g/L, 0.243 g/L and 5.29, respectively. New batch fermentation as an external validation was used to check the models, and the results suggested that the predictive capacity of the models for the glutamate fermentation process was good.

Combination of fractional order derivative and memory-based learning algorithm to improve the estimation accuracy of soil organic matter by visible and near-infrared spectroscopy

Mapping soil properties, using geostatistical methods in support of precision agriculture and related activities, requires a large number of samples. To reduce soil sampling and measurement time and cost, a combination of field proximal soil sensors was used to predict and map laboratory-measured soil properties in a 3.4-ha pasture field in southeastern Brazil. Sensor soil properties were measured in situ on a 10 × 10-m dense grid (377 samples) using apparent electrical conductivity meters, apparent magnetic susceptibility meter, gamma-ray spectrometer, water content reflectometer, cone penetrometer, and portable X-ray fluorescence spectrometer (pXRF). Soil samples were collected on a 20 × 20-m thin grid (105 samples) and analyzed in the laboratory for organic C, sum of bases, cation exchange capacity, clay content, soil volumetric moisture, and bulk density. Another 25 samples collected throughout the area were also analyzed for the same soil properties and used for independent validation of models and maps. To test whether the combination of sensors enhances soil property predictions, stepwise multiple linear regression (MLR) models of the laboratory soil properties were derived using individual sensor covariate data versus combined sensor data—except for the pXRF data, which were evaluated separately. Then, to test whether a denser grid sample boosted by sensor-based soil property predictions enhances soil property maps, ordinary kriging of the laboratory-measured soil properties from the thin grid was compared to ordinary kriging of the sensor-based predictions from the dense grid, and ordinary cokriging of the laboratory properties aided by sensor covariate data. The combination of multiple soil sensors improved the MLR predictions for all soil properties relative to single sensors. The pXRF data produced the best MLR predictions for organic C content, clay content, and bulk density, standing out as the best single sensor for soil property prediction, whereas the other sensors combined outperformed the pXRF sensor for the sum of bases, cation exchange capacity, and soil volumetric moisture, based on independent validation. Ordinary kriging of sensor-based predictions outperformed the other interpolation approaches for all soil properties, except organic C content, based on validation results. Thus, combining soil sensors, and using sensor-based soil property predictions to increase the sample size and spatial coverage, leads to more detailed and accurate soil property maps.

This study used visible/near-infrared hyperspectral imaging (HSI) technology combined with chemometric methods to assess the freshness of pearl gentian grouper. The partial least square discrimination analysis (PLS-DA) and competitive adaptive reweighted sampling-PLS-DA (CARS-PLS-DA) models were used to classify fresh, refrigerated, and frozen–thawed fish. The PLS-DA model achieved better classification of fresh, refrigerated, and frozen–thawed fish with the accuracy of 100%, 96.43%, and 96.43%, respectively. Further, the PLS regression (PLSR) and CARS-PLS regression (CARS-PLSR) models were used to predict the storage time of fish under different storage conditions, and the prediction accuracy was assessed using the prediction correlation coefficients (Rp2), root mean squared error of prediction (RMSEP), and residual predictive deviation (RPD). For the prediction of storage time, the CARS-PLS model presented the better result of room temperature (Rp2 = 0.948, RMSEP = 0.255, RPD = 4.380) and refrigeration (Rp2 = 0.9319, RMSEP = 1.188, RPD = 3.857), while the better prediction of freeze was by obtained by the PLSR model (Rp2 = 0.9250, RMSEP = 2.910, RPD = 3.469). Finally, the visualization of storage time based on the PLSR model under different storage conditions were realized. This study confirmed the potential of HSI as a rapid and non-invasive technique to identify fish freshness.

<p>Soil properties could be assessed with reflectance spectroscopy (soil spectroscopy, SS) in vis-NIR region (400-2500 nm) through absorption features found in soil spectra. A high spectral resolution (up to 1 nm) drives to high dimensional and multicollinear data. This issue is usually addressed prior to modelling with feature extraction methods such as Principal Component Analysis, or embedded methods such as Partial Least Squares Regression (PLSR). Feature Selection (FS) wrapper methods are promising dimensionality reduction approaches barely used in SS. The objective of this study was two-fold: i) evaluate the performance of FS wrapper methods built from Random Forest (RF) algorithm to predict soil organic matter (SOM), clay and carbonates using laboratory spectroscopy, ii) test the performance of FS methods for dimensionality reduction in SS. The reflectance of 100 soil samples from Sierra de las Nieves National Park (Spain), was measured under laboratory conditions using an ASD FieldSpec Pro JR. A spectral preprocessing method, Continumm Removal (CR), was applied to raw spectra. The RF wrapper considered two different feature searching approaches: Sequential Forward Selection (SFS) and Sequential Flotant Forward Selection (SFFS). The performance of RF with FS (RF-FS) was compared to that of Partial Least Squares Regression (PLSR) and RF (without FS). Models were evaluated with R-squared, root mean squared error (RMSE) and ratio of prediction to deviation (RPD).</p><p>RF-FS models outperformed PLSR and RF models for the three SAP. RF-FS best models had a RPD of 2.19 for SOM, 1.64 for carbonates and 1.52 for clay, whereas PLSR models had RPD values of 1.59, 1.22 and 1.3, and RF 1.38, 1.23 and 1.23 for SOM, carbonates, and clay, respectively. Therefore, FS was useful in obtaining models with improved accuracy by reducing redundant features and avoiding multicollinearity (Hughes effect). The application of FS wrapper methods reduced the number of features in the RF-FS models to less than 1% of the starting features. Features were selected across all spectra from SOM and clay, and around 900, 1900 and 2350 nm for carbonates. This research, thus, shows an alternative to different feature extraction approaches for modelling soil properties based on FS methods and machine learning.</p>

The development of rapid, accurate, and cost-effective methods to determine forest soil properties such as visible and near infrared (Vis-NIR) spectroscopy is important for sustainable land management. The main objective of this study was to assess the suitability of Vis-NIR spectroscopy coupled with partial least squares regression (PLSR) to determine some forest soil properties such as organic carbon (SOC), total nitrogen (TN), pH, and soil texture (sand, silt, and clay) for a representative forest area in southern Italy. Soil samples (0–20 cm depth) were collected at 267 locations, oven-dried and passed through a 2-mm sieve, and analyzed for some chemical and physical soil properties using conventional laboratory methods. Vis-NIR reflectance of each soil sample was measured in laboratory under artificial light using an ASD FieldSpec IV 350–2500 nm spectroradiometer. Partial least squares regression (PLSR) was used to develop a calibration model for SOC, TN, pH, sand, silt, and clay. Samples were split into a calibration set (187 samples) to develop the models and a validation set (80 samples) to assess the prediction accuracy of the calibration models. Results showed a good agreement between measured and predicted values with high R 2 and low root mean square error (RMSE) values. Model validation using independent data was satisfactory for all the studied soil properties. Finally, findings confirmed that laboratory Vis–NIR spectroscopy has the potential to be a non-destructive and cost-effective tool for rapid determination of many soil properties. The spectral data collected in this study could contribute to build a regional soil spectral library to be used advantageously in support to soil survey in other areas of the Calabria region (southern Italy).