Soil cation exchange capacity (CEC) is an important measure of soil fertility, owing to the large number of plant essential nutrients which are taken up by plants as cations. However, current methods ...for evaluating soil CEC are arduous and require analysis in a laboratory. Portable X-ray fluorescence (PXRF) spectrometry is a proximal sensing technique which provides elemental data in-situ, in seconds. This study examined the potential of using PXRF for soil CEC prediction by evaluating 450 soil samples from California and Nebraska, USA representing a wide variety of soil textures found in active farm fields. Multiple linear regression was applied to a modeling dataset to establish the relationship between lab-determined CEC and PXRF elemental data. A second model also included auxiliary input data (soil clay, pH, organic matter) as potential modeling variables. Both models were shown to perform similarly, with the auxiliary input model providing slightly higher R2 (0.926 vs. 0.908) and slightly lower RMSEs (2.236 vs. 2.498) compared to pure elemental data models. Independent validation datasets were compelling for both pure elemental models (0.904) and auxiliary input models (0.953). Summarily, PXRF was able to predict soil CEC accurately, thereby minimizing the need for lab-based CEC data for many applications.
•X-ray fluorescence (XRF) elemental concentrations were correlated to soil CEC.•Elements germane to soil CEC were identified.•Multiple linear regression models successfully predicted soil CEC from PXRF data.
•A setup containing a smartphone and a dark chamber was used for predicting soil texture.•Soil images were acquired using the setup and analyzed via computer vision, RF, and CNN models.•Both RF and ...CNN showed high prediction accuracy for clay and sand, with moderate prediction accuracy for silt.•Image-extracted color features showed the maximum influence on the RF model performance.•An Android app using the calibrated CNN model was able to predicted soil textural values with satisfactory accuracy.
The rapid and non-invasive prediction of soil sand, silt, and clay is becoming increasingly attractive given the laborious nature of traditional soil textural analysis. This study proposed a novel and cheap setup comprising a smartphone, a custom-made dark chamber, and a smartphone application for predicting soil texture of the dried, ground, and sieved samples. The image acquisition system was used to capture triplicate images from 90 mineral soil samples, representing a wide textural variability from sand to clay. Local features, color features, and texture features were extracted from the cropped images and subsequently used in different combinations to predict laboratory-measured clay, silt, and sand via random forest (RF) and convolutional neural network (CNN) algorithms. Results indicated high prediction accuracy for clay (R2 = 0.97–0.98) and sand (R2 = 0.96–0.98) and moderate prediction accuracy for silt (R2 = 0.62–0.75) using both algorithms. Color features outperformed all other image-extracted features and showed the maximum influence on RF model performance. The better performance of the color features can be attributed to the color features of mineral matter and soil organic matter (SOM). An Android-based smartphone application based on the calibrated CNN model was able to predict and return soil textural values. These results exhibited the potential of the proposed system as a proximal sensor for rapid, cost-effective, and eco-friendly soil textural analysis using computer-vision and deep learning. More research is warranted to augment the setup design, develop a standalone mobile application, and measure the impacts of soil moisture and high SOM on the model prediction performance to extend the approach for on-site prediction of soil texture.
Knowing the geographical origins of chili papers produced in specific areas is crucial because the geographical origins of various varieties of chili powder have a significant impact on their quality ...and price. In this research, for the first time, NIR (near-infrared) spectroscopy was used for the identification and classification of the geographical origin of chili powder of 6 different varieties, combining the method of PCA (principal component analysis) to extract relevant spectral features from the spectral data and segregate visible cluster trends, SIMCA (soft independent modeling of class analogy) statistically based classification model, and the four machine learning (ML) classifiers, including K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM), were applied for supervised classification. It was found that the SVM classifier, with a
C
value of 4013.0 and γ of 0.04125, delivered the highest cross-validation accuracy of 98.41% and prediction accuracy of 97.22%. The optimization process, guided by a detailed 3D contour plot, led to a model that not only generalized well but also offered remarkable precision, as confirmed by confusion matrices. The classification accuracy of the SIMCA model was 94.04% for the calibration set and 84.74% for the prediction set. The nonlinear SVM technique of classification outperformed the linear SIMCA model and other ML models. In general, the results indicated that chili powder from various geographic origins could be discriminated by the use of NIR spectroscopy combined with the SVM model quickly, nondestructively, and reliably.
Indian soils are inherently poor in quality due to the warm climate and erosion. Conversion of land uses like forests to croplands and faulty management practices in croplands further cause soil ...degradation. This study aimed to understand the extent of these impacts in a small representative part of eastern India, covering Himalayan terai and nearing alluvial plains. Soils were collected from (i) forests, (ii) croplands (under agricultural practices for more than 50-60 years) and (iii) converted lands (converted from forests to croplands or tea gardens over the past 15-20 years). Different soil quality indicators were assessed and soil quality index (SQI) was generated to integrate, scale and allot a single value per soil. Results indicated that continuous organic matter deposition and no disturbances consequence the highest presence of soil carbon pools, greater aggregation and maximum microbial dynamics in forest soils whereas high application of straight fertilizers caused the highest available nitrogen and phosphorus in cropland soils. The SQI scorebook indicated the best soil quality under forests (xMACRON 0.532), followed by soils of converted land (xMACRON 0.432) and cropland (xMACRON 0.301). Comparison of the SQI spatial distribution with land use and land cover confirmed the outcome. Possibly practices like excessive tillage, high cropping intensity, no legume in crop rotations, cultivation of heavy feeder crops caused degraded soil quality in croplands. This study presented an example of soil quality degradation in India due to land use change and faulty management practices. Such soil degradation on a larger scale may affect future food security.
•Smartphone-images were used to classify 12 rice diseases and deficiency symptoms.•Image processing and model optimization were used to develop an android application.•MobileNet V2 appeared as the ...best model for the developed smartphone application.•The “Rice Disease Detector” performed well for multiple disease occurrences in a plant.
The earliest detection of plant disease is the primary concern of the farming community. The availability of advanced digital cameras and smartphones with improved image acquisition modes and deep learning methods like convolutional neural networks (CNN) can detect plant disease with high accuracy. This study used 2259 smartphone images of various rice (Oryza sativa) plant parts under various classes and 250 real-time validation images for classifying 12 rice diseases and nutrient deficiency symptoms. Different image segmentation techniques like foreground extraction were used to segment affected portions. Additionally, optimization of models and procedures to use them in smartphones with offline working capabilities have been described. Furthermore, a dynamic framework has been developed and demonstrated where the server trains on unseen image data to improve the classification performance and updates the model on falling below a certain threshold level. A comparison was made across different models used for image classification with many supporting metrics to select the best model for transfer learning. The top four performing models were DenseNet201, Xception, MobileNetV2, and ResNet50, with validation accuracies of 0.9803, 0.9778, 0.9756, and 0.9718, respectively. The ResNet50 model was found to best among all for cloud architectures, while MobileNetV2 appeared as the best model for the smartphone application. Finally, the android application, “Rice Disease Detector” compiled with the MobileNetV2 model was tested for multiple disease occurrences in a single capture. More research is warranted to test the application for smartphones with variable configurations.
Display omitted
Display omitted
•Elemental data from PXRF scans predicted water salinity with reasonable accuracy.•PXRF Cl was most useful in assessing water salinity.•PXRF shows promise for rapid assessment of ...metal/pollutant laden waters.
Rapid, in-situ elemental water analysis would be an invaluable tool in studying polluted and/or salt-impacted waters. Analysis of water salinity has commonly used electrical conductance (EC); however, the identity of the elements responsible for the salinity are not revealed using EC. Several studies have established the viability of using portable X-ray fluorescence (PXRF) spectrometry for elemental data analysis of soil, sediment, and other matrices. However, the accuracy of PXRF is known to be affected while scanning moisture-laden soil samples. This study used PXRF elemental data in water samples to predict water EC. A total of 256 water samples, from 10 different countries were collected and analyzed via PXRF, inductively coupled plasma atomic emission spectroscopy (ICP-AES), and a digital salinity bridge. The PXRF detected some elements more effectively than others, but overall results indicated that PXRF can successfully predict water EC via quantifying Cl in water samples (validation R2 and RMSE of 0.77 and 0.95logμScm−1, respectively). The findings of this study elucidated the potential of PXRF for future analysis of pollutant and/or metal contaminated waters.
•Munsell color was accurately predicted by Nix Pro color sensor.•Mean color difference (ΔE*ab) achieved 2.9, indicating the Nix Pro accuracy.•Results achieved overall accuracy and Kappa index of ...0.93.•Using a dirty, old Munsell color chart reduced the accuracy of predictions.•Soil color determined by human eye was compared to Nix Pro color predictions.
Soil color has historically drawn humans’ attention, although its definition is somewhat subjective. It is correlated with several soil attributes and it allows for inferences about several soil aspects. New proximal sensors, such as the Nix™ Pro color sensor, can determine soil color values, but its correlation with the widely used Munsell soil color chart (MSCC) has yet to be investigated. This work aimed to train machine learning models using the Random Forest (RF) algorithm to predict each notation of MSCC chips from data extracted by the Nix™ Pro sensor, test the model’s accuracy by evaluating whether it can identify MSCC chips using a brand-new and a dirty MSCC, and compare model predictions with soil color classifications made by the human eye. Additionally, MSCC data obtained via Nix™ was compared to Munsell renotation data to assess the color detection accuracy of the sensor. Prediction models were calibrated by scanning every MSCC chip (437 in total) in triplicate. All validation samples were excluded from model calibration. Accuracy of the predictions of MSCC notation reached overall accuracy and Kappa index values of 0.93 for the brand-new MSCC and of 0.70 for the dirty MSCC. Soil color classification by human eye had little agreement with the predicted MSCC notation, as expected due to the variable conditions affecting soil color conventional determination in the field. Color difference was calculated by the Euclidian distance (ΔE*ab) between three color stimuli in the CIELAB color space. The mean ΔE*ab between Nix™-provided data and renotation data was 2.9, demonstrating high color detection accuracy. The Nix™ Pro color sensor allows for assessment of accurate color data. When applied together with machine learning algorithms, Nix™ Pro provides a reliable determination of soil color classification equivalent to MSCC in an easily reproducible, rapid, inexpensive and non-subjective way.
Portable X‐ray fluorescence (PXRF) spectrometry is a proximal sensing technique whereby low‐power X‐rays are used to make elemental determinations in soils. The technique is rapid, portable, and ...provides multi‐elemental analysis with results generally comparable to traditional laboratory‐based techniques. Elemental data from PXRF can then be either used directly for soil parameter assessment (e.g., total Ca, total Fe) or as a proxy for predicting other soil parameters of interest (e.g., soil cation‐exchange capacity CEC, soil reaction, soil salinity) via simple or multiple linear regression. Importantly, PXRF does have some limitations that must be considered in the context of soil analysis. Those notwithstanding, PXRF has proven effective in numerous, agronomic, pedological, and environmental quality assessment applications.
In this study, a novel chromotropic acid-based color development method was proposed for quick estimation of soil nitrate (NO3−). The method utilized a 3D printed device integrated with the rear-end ...camera of a smartphone and a stand-alone application called SMART NP. By analyzing the mean Value (V) component of the sample’s image, the SMART NP provides instant predictions of soil NO3− levels. The limit of detection was calculated as 0.1 mg L−1 with a sensitivity of 0.26 mg L−1. The device showed a % bias of 0.9% and a precision of 1.95%, indicating its reliability. Additionally, the device-predicted soil NO3− data, combined with kriging interpolation, showcased spatial variability in soil NO3− levels at the regional level. The study employed a Gaussian model of variogram for kriging, and the high Nugget/Sill ratio indicated low spatial autocorrelation, emphasizing the impact of management factors on the spatial distribution of soil NO3− content in the study area. Overall, the imaging device, along with geostatistical interpolation, provided a comprehensive solution for the rapid assessment of spatial variability in soil NO3−content.
A common theme emerging from nutrient omission trials conducted extensively across sub-Saharan Africa has been the large variability in yield response to applied nutrients. Yet, little is known about ...the factors associated with this variability. Therefore, the objectives of this review were to (1) synthesize the available data from nutrient omission trials of maize and rice and quantify spatial variability in yield responses and the probability of attaining specified yield targets; (2) identify the genotypic, environmental, and nutrient management factors associated with yield responses; and (3) provide insights and actionable information for setting priorities for future nutrient management research and development. Here, we provide distributions and expected values of yield response and agronomic efficiencies of nitrogen (N), phosphorus (P), and potassium (K) to serve as benchmarks for improving fertilizer recommendations. We also show that with the combined application of NPK fertilizer, rain-fed maize and rice yields can be raised from the current yields of ~2 Mg ha
−1
to 4 Mg ha
−1
. Yield responses to N, P, and K were in the ratio of 11:6:1 in maize and 13:7:5 in rice, probably arising from differences in uptake requirements and growing environments. K yield responses were 2–3 times higher in rice than in maize. Explorative analysis using machine learning algorithms provided further insights into the possible interaction of agroecology, soil type, and exchangeable cations on the spatial variability in yield responses to N, P, and K in maize and rice. We recommend future research to address site-specific interactions between the applied and indigenous soil nutrients, focusing on optimizing application rates of K, other macronutrients and micronutrients for sustainable intensification of maize and rice production. This study highlights the critical need for balanced fertilization to optimize the productivity and nutrient use efficiencies in rice and maize production in smallholder farming systems in SSA.
PDF
