Reflectance Spectra Of Soil Samples Research Articles

Estimation of heavy metals using deep neural network with visible and infrared spectroscopy of soil

Heavy metal contamination in soil disturbs the chemical, biological, and physical soil conditions and adversely affects the health of living organisms. Visible and near-infrared spectroscopy (VNIRS) shows a potential feasibility for estimating heavy metal elements in soil. Moreover, deep learning models have been shown to successfully deal with complex multi-dimensional and multivariate nonlinear data. Thus, this study implemented a deep learning method on reflectance spectra of soil samples to estimate heavy metal concentrations. A convolutional neural network (CNN) was adopted to estimate arsenic (As), copper (Cu), and lead (Pb) concentrations using measured soil reflectance. In addition, a convolutional autoencoder was utilized as a joint method with the CNN for dimensionality reduction of the reflectance spectra. Furthermore, artificial neural network (ANN) and random forest regression (RFR) models were built for heavy metal estimation. Principal component analysis was utilized for dimensionality reduction of the ANN and RFR models. Among these models, the CNN model with convolutional autoencoder showed the highest accuracies for As, Cu, and Pb estimates, having R2 values of 0.86, 0.74, and 0.82, respectively. The convolutional autoencoder disentangled the relevant features of multiple heavy metal elements and delivered robust features to the CNN model, thereby generating relatively accurate estimates.

The application of proximal visible and near-infrared spectroscopy to estimate soil organic matter on the Triffa Plain of Morocco

Soil organic matter (SOM) is a fundamental soil constituent. The estimation of this parameter in the laboratory using the classical method is complex time-consuming and requires the use of chemical reagents. The objectives of this study were to assess the accuracy of two laboratory measurement setups of the VIS-NIR spectroscopy in estimating SOM content and determine the important spectral bands in the SOM estimation model. A total of 115 soil samples were collected from the non-root zone (0–20 cm) of soil in the study area of the Triffa Plain and then analysed for SOM in the laboratory by the Walkley–Black method. The reflectance spectra of soil samples were measured by two protocols, Contact Probe (CP) and Pistol Grip (PG)) of the ASD spectroradiometer (350–2500 nm) in the laboratory. Partial least squares regression (PLSR) was used to develop the prediction models. The results of coefficient of determination (R2) and the root mean square error (RMSE) showed that the pistol grip offers reasonable accuracy with an R2 = 0.93 and RMSE = 0.13 compared to the contact probe protocol with an R2 = 0.85 and RMSE = 0.19. The near-Infrared range were more accurate than those in the visible range for predicting SOM using the both setups (CP and PG). The significant wavelengths contributing to the prediction of SOM for (PG) setup were at: 424, 597, 1432, 1484, 1830,1920, 2200, 2357 and 2430 nm, while were at 433, 587, 1380, 1431, 1929, 2200 and 2345 nm for (CP) setup.

Predicting nickel concentration in soil using reflectance spectroscopy associated with organic matter and clay minerals

Heavy metal contamination in soil has become a serious environmental problem. Visible and near-infrared reflectance spectroscopy (VNIRS) is recognized as an alternative to rapidly predict heavy metal concentration in soil. The correlation between soil spectrally active constituents and heavy metals provides a mechanism for the prediction. Generally, the entire VNIR spectral region (VNIR-SR) without discrimination is used. Considering the adsorption and retention of heavy metals on soil spectrally active constituents, a method was proposed to predict heavy metal concentration in soil using VNIRS. Organic matter and clay minerals have strong sorption and retention for Ni in soil. Therefore, the spectral bands associated with organic matter and clay minerals were used to predict nickel (Ni) concentration to validate the proposed method. In this study, two sets of reflectance spectra of soil samples collected in Chenzhou and Hengyang, Hunan Province, China were used. The prediction model was calibrated with a combination of genetic algorithm and partial least squares regression (GA-PLSR). In Chenzhou, the ratio of prediction to deviation (RPD) and the coefficient of determination (R2) were improved from 1.566 and 0.577 to 2.139 and 0.773 by the prediction using the spectral bands associated with organic matter and clay minerals compared with the prediction using the entire VNIR-SR. The RPD and R2 were improved from 1.805 and 0.672 to 3.144 and 0.892 by the proposed method in Hengyang. To further validate the proposed method, the soil samples from Chenzhou were reallocated to calibration and validation sets according to Ni concentration. The RPD and R2 were improved from 1.193 and 0.267 to 2.396 and 0.818 by using the spectral bands associated with organic matter and clay minerals. The results indicate that the proposed method is effective in predicting Ni concentration and has the potential to predict other heavy metals in soil.

Linear Regression between CIE-Lab Color Parameters and Organic Matter in Soils of Tea Plantations

To quantify the relationship between the soil organic matter and color parameters using the CIE-Lab system, 62 soil samples (0–10 cm, Ferralic Acrisols) from tea plantations were collected from southern China. After air-drying and sieving, numerical color information and reflectance spectra of soil samples were measured under laboratory conditions using an UltraScan VIS (HunterLab) spectrophotometer equipped with CIE-Lab color models. We found that soil total organic carbon (TOC) and nitrogen (TN) contents were negatively correlated with the L* value (lightness) (r = –0.84 and –0.80, respectively), a* value (correlation coefficient r = –0.51 and –0.46, respectively) and b* value (r = –0.76 and –0.70, respectively). There were also linear regressions between TOC and TN contents with the L* value and b* value. Results showed that color parameters from a spectrophotometer equipped with CIE-Lab color models can predict TOC contents well for soils in tea plantations. The linear regression model between color values and soil organic carbon contents showed it can be used as a rapid, cost-effective method to evaluate content of soil organic matter in Chinese tea plantations.

Spectroscopy-Based Soil Organic Matter Estimation in Brown Forest Soil Areas of the Shandong Peninsula, China

Soil organic matter (SOM) is important for plant growth and production. Conventional analyses of SOM are expensive and time consuming. Hyperspectral remote sensing is an alternative approach for SOM estimation. In this study, the diffuse reflectance spectra of soil samples from Qixia City, the Shandong Peninsula, China, were measured with an ASD FieldSpec 3 portable object spectrometer (Analytical Spectral Devices Inc., Boulder, USA). Raw spectral reflectance data were transformed using four methods: nine points weighted moving average (NWMA), NWMA with first derivative (NWMA + FD), NWMA with standard normal variate (NWMA + SNV), and NWMA with min-max standardization (NWMA + MS). These data were analyzed and correlated with SOM content. The evaluation model was established using support vector machine regression (SVM) with sensitive wavelengths. The results showed that NWMA + FD was the best of the four pretreatment methods. The sensitive wavelengths based on NWMA + FD were 917, 991, 1 007, 1 996, and 2 267 nm. The SVM model established with the above-mentioned five sensitive wavelengths was significant (R2 = 0.875, root mean square error (RMSE) = 0.107 g kg−1 for calibration set; R2 = 0.853, RMSE = 0.097 g kg−1 for validation set). The results indicate that hyperspectral remote sensing can quickly and accurately predict SOM content in the brown forest soil areas of the Shandong Peninsula. This is a novel approach for rapid monitoring and accurate diagnosis of brown forest soil nutrients.

Targeting and mapping expansive soils (Loiret, France): geometrical analysis of laboratory soil spectra in the short-wave infrared domain (1100–2500 nm)

Abstract Short-wave infrared (SWIR: 1100–2500 nm) reflectance spectra of soil samples, acquired under laboratory-controlled conditions, were used to estimate the swelling potentials of the soils and to use SWIR spectroscopy to improve a part of the swelling-risk map of France. A total of 332 samples were collected to the west of Orléans (France) in various geological formations and swelling-risk areas and along two (eastern and western) transects with different samples spacings. Comparisons between the swelling potentials of the soils and the swelling-risk areas of the map exhibit good correlation in the south of the sampling area; however, there are several inconsistencies in the north of the study area that highlight the necessity of locally redrawing the accepted swelling-risk map of France. The sampling interval (approximately 260 m) along the eastern transect was too sparse and does not appear to have effectively captured the smallest and/or isolated lithologies. The sampling interval along the western transect (approximately 50 m) revealed the presence of an unmapped swelling-potential zone, which was confirmed by several soil samples. The sample interval along the western transect appears to be more suitable for mapping the smallest lithologies. The presence of several samples in close proximity that exhibit the same swelling potential is a robust indication of the presence of a zone with constant swelling potential. A new swelling-risk map of the sampling area was produced based on the soil samples. The map produced by interpolation did not permit the representation of discrete lithological units, introduced spurious swelling-risk zones that however could be representative of tillage in agricultural zones. More samples are therefore needed to produce a reliable map on the scale of the sampling area. According to swelling potential uncertainty related to soil sampling and soil treatment, spectroscopy-based approach proposed here cannot be used to replace the existing swelling-risk map of France. This method permits however the rapid and low-cost estimation of the swell potentials of a large number of samples, which could be used at regional-scale to target areas where doubt remains or where infrastructure damages attributed to swelling behavior are known. At local scale, soil samples need to be properly and laboratory treated to accurately produce local revised and detailed swelling-risk maps.

Soil organic carbon content estimation with laboratory-based visible-near-infrared reflectance spectroscopy: feature selection.

This study, with Yixing (Jiangsu Province, China) and Honghu (Hubei Province, China) as study areas, aimed to compare the successive projection algorithm (SPA) and the genetic algorithm (GA) in spectral feature selection for estimating soil organic carbon (SOC) contents with visible-near-infrared (Vis-NIR) reflectance spectroscopy and further to assess whether the spectral features selected from one site could be applied to another site. The SOC content and Vis-NIR reflectance spectra of soil samples were measured in the laboratory. Savitzky-Golay smoothing and log10(1/R) (R is reflectance) were used for spectral preprocessing. The reflectance spectra were resampled using different spacing intervals ranging from 2 to 10 nm. Then, SPA and GA were conducted for selecting the spectral features of SOC. Partial least square regression (PLSR) with full-spectrum PLSR and the spectral features selected by SPA (SPA-PLSR) and GA (GA-PLSR) were calibrated and validated using independent datasets, respectively. Moreover, the spectral features selected from one study area were applied to another area. Study results showed that, for the two study areas, the SPA-PLSR and GA-PLSR improved estimation accuracies and reduced spectral variables compared with the full spectrum PLSR in estimating SOC contents; GA-PLSR obtained better estimation results than SPA-PLSR, whereas SPA was simpler than GA, and the spectral features selected from Yixing could be well applied to Honghu, but not the reverse. These results indicated that the SPA and GA could reduce the spectral variables and improve the performance of PLSR model and that GA performed better than SPA in estimating SOC contents. However, SPA is simpler and time-saving compared with GA in selecting the spectral features of SOC. The spectral features selected from one dataset could be applied to a target dataset when the dataset contains sufficient information adequately describing the variability of samples of the target dataset.

Predicting salinity in tomato using soil reflectance spectra

Soil salinity is one of the most common soil degradation processes, found particularly in both arid and semi-arid areas. Salt (Cl)- and sodium (Na)-affected soils impact vegetation and plant communities. Under these conditions, soil salinity can serve as an indicator of vegetation salinity. In this study, we explored whether spectroscopy could quantitatively assess foliar Cl and Na concentration as indicators to assess salinity in tomato plants. Reflectance spectra of soil samples were obtained in the 400–2500 nm region using a hyperspectral radiometer. The relationship between the Na and Cl contents of tomato plants growing in various saline environments and soil spectral reflectance was determined using partial least squares regression. The Cl-content model was more accurate for determining leaf salinity (R 2 = 0.92, root mean square error of prediction (RMSEP) = 0.2%) than the Na-content model (R 2 = 0.87, RMSEP = 0.6%). We conclude that reflectance spectroscopy is potentially useful for characterizing the key properties of salinity in growing vegetation and assessing its salt quality. The results of this study can serve as a starting point in precision agriculture for salinity measurements in tomato fields and could be further upgraded for use by airborne/satellite remote-sensing modes.

Testing of the Region of Murcia soils by near infrared diffuse reflectance spectroscopy and chemometrics

A partial least squares near infrared (PLS-NIR) method has been developed for the determination of several physicochemical parameters in soils from different locations of the Region of Murcia. The method was based on the proper chemometric treatment of diffuse reflectance spectra of soil samples. Reflectance spectra were scanned from samples stored in glass vials in the NIR region between 800 and 2600 nm, averaging 36 scans per spectrum at a resolution of 8 cm −1. Models were built using reference data of 39 samples selected from a dendrogram obtained after hierarchical cluster analysis of NIR spectra of soils and prediction parameters were established from a validation set of 109 additional samples of the same area not considered to build the model. Organic matter, CaCO 3, pH, electrical conductivity (EC), together with several trace metals as Cr, Co, Ni, Cu, Zn, As, Se, Cd and Tl, were employed as characteristic parameters of the soils under study, and found results evidenced that PLS-NIR provides a valuable tool for screening purposes providing residual predictive deviations which ranged from 0.9 to 1.5 as a function of the considered parameter.

Assessing Reclamation Levels of Coastal Saline Lands with Integrated Stepwise Discriminant Analysis and Laboratory Hyperspectral Data

ABSTRACT At different times over the past 30 years in Zhejiang Province, China, the coastal tidelands have been successively enclosed and reclaimed for agricultural land use. The purpose of this work was to evaluate whether laboratory hyperspectral data might be used to estimate the physicochemical characteristics of these reclaimed saline soils. A coastal region of Shangyu City (Zhejiang Province), which was grouped into four subzones according to reclamation history, was used as the study area, and soil samples were collected in each subzone. Physicochemical analyses showed that the soils were characterized by high electrical conductivity and sand content with low organic matter; the longer the saline lands had been reclaimed, the lower were the electrical conductivity and sand content and the higher the organic matter content. These changing trends of soil chemical and physical properties were found in laboratory reflectance spectra of soil samples and their first-order derivative curves. Stepwise discriminant analysis (SDA) identified six salient spectral bands at 488, 530, 670, 880, 1400, and 1900 nm. Using derived discriminant functions for saline lands with different historical years of reclamation, classification revealed an overall accuracy from a self-test of 86.6% and from cross-validation of 89.3%. Therefore, as opposed to time-consuming field investigations, this study suggested that remotely sensed hyperspectral data could serve as a promising measure to assess the reclamation levels of coastal saline lands.

Some theoretical properties of the O‐PLS method

AbstractThe objective of this paper is to present new properties of the orthogonal projections to latent structures (O‐PLS) method developed by Trygg and Wold (J. Chemometrics 2002; 16: 119–128). The original orthogonal signal correction (OSC) filter of Wold et al. (Chemometrics Intell. Lab. Syst. 1998; 44: 175–185) removes systematic variation from X that is unrelated to Y. O‐PLS is a more restrictive OSC filter. O‐PLS removes only systematic variation in X explained in each PLS component that is not correlated with Y. O‐PLS is a slight modification of the NIPALS PLS algorithm, which should make O‐PLS a generally applicable preprocessing and filtering method. The computation of the O‐PLS components under the constraint of being correlated with one PLS component imposes particular properties on the space spanned by the O‐PLS components. This paper is divided into two main sections. First we give an application of O‐PLS on near‐infrared reflectance spectra of soil samples, showing some graphical properties. Then we give the mathematical justifications of these properties. Copyright © 2004 John Wiley & Sons, Ltd.

Multivariate Analysis of Laboratory Spectra for the Assessment of Soil Development and Soil Degradation in the Southern Apennines (Italy)

Multivariate statistical analysis was applied to high resolution laboratory reflectance spectra of soil samples reflecting the range of soil development levels in a test site in the southern Apennines (Fortore beneventano) in Italy. Principal component analysis was used to identify the dominant characteristics of soil spectra and cluster analysis to group samples into homogeneous classes. The analysis was applied stepwise (1) to a minimum number of reflectance values, representative of the general shape and curvature of soil spectra, and (2) to the normalized amplitude (depth) of characteristic absorption bands. In the first step analysis, it was found that the brightness and the visible near-infrared (VIS-NIR) slope of the spectra are important spectral characteristics to discriminate soil development, also in relation to the underlying lithology. The principal intercorrelation between the brightness, the VIS-NIR spectral slope, and various soil properties, such as calcium carbonate content, organic matter, and texture, could be also explained. In the second step, it was demonstrated that different types of soil included in the homogeneous groups, resulting from the first step multivariate analysis, could be further separated, mainly as function of the iron oxide features and the two main water and hydroxyl absorption bands at 1400 nm and 1900 nm. These results suggest that our method is suitable for the evaluation and interpretation of remote sensing data for discriminating different soil development and degradation levels of large parts of the Apennines, which are similar to the test area. The brightness and the VIS-NIR slope components derived in the first step multivariate analysis may be retrievable from common multispectral systems such as Landsat-TM. The application of the second step multivariate analysis, however, would require the use of hyperspectral systems capable of retrieving the absorption features. In both cases rigorous atmospheric correction including the removal of the effects of relief illumination is a prerequisite.

Spectral discrimination of contact metamorphic zones and its potential for mineral exploration, province of Salamanca, Spain

In the mineralized contact aureole of Los Santos, Spain, contact metamorphic grade of rocks and the mineralogical composition of overlying soils correlate systematically. From the exocontact towards the granite intrusion the proportions of kaolinite, illite and free iron in soils increase, while the proportions of chlorite and smectite decrease. Laboratory reflectance spectra of soil samples demonstrate that with the increase of the proportion of kaolinite, illite, and free iron and the decrease of chlorite, ratios of Band 2 / Band 3, Band 2 / Band 5, and Band 2 / Band 7 decrease while ratios of Band 3 / Band 4, Band 4 / Band 7, and Band 5 / Band 7 increase. Soil reflectance is significantly affected if more than 1.5 % organic carbon is present. A procedure is presented that enables mapping of soils overlying contact metamorphic rocks in Thematic Mapper (TM) imagery, using these variations. In the TM image, pixels that correspond with kaolinitic soils are identified by elimination of other pixels during a process of stepwise masking various band ratio images. The first step was to eliminate the effect of vegetation and organic material, by masking high values of the ratio of ((TM5 / TM7) / (TM3 /TM4)]. The second step was to eliminate chloritic soils by masking the ratio of TM2 / TM7. The third step was to isolate the most kaolinitic and limonitic soils using the ratio of TM3 / TM4. Finally, a maximum likelihood classification was performed on original band ratio images, using the result of the masking as training set, in order to optimize the classification results.

Detectability of minerals on desert alluvial fans using reflectance spectra

We analyzed the visible and near‐infrared reflectance spectra of soil samples collected from desert alluvial and colluvial surfaces in the Cuprite mining district, Nevada. These surfaces are downslope from hydrothermally altered volcanic rocks that contain spectrally characteristic minerals such as alunite and kaolinite. Coarse fractions of the soils on the alluvial fans are mineralogically variable and express the upslope lithologies; fine fractions are remarkably similar mineralogically and spectrally in all samples because of dilution of local mineral components by regionally derived windblown dust. Theoretical models for spectral mixing and for particle‐size effects were used to model the observed spectral variations. Diagnostic mineral absorption bands in the spectra of fan materials were enhanced by computationally removing the spectrum of the homogenous fine‐soil component. Detection limits for minerals were found to be controlled by the sensitivity of the instrument, the spectral contrast between the mineral and the background, and the relative brightnesses of the mineral and the background. These results show that spectral mixing models are useful for analyzing data with high spectral resolution obtained by field and aircraft spectrometers.