Soil spectroscopy with the use of chemometrics, machine learning and pre-processing techniques in soil diagnosis: Recent advances–A review

Barra, Issam; Haefele, Stephan M.; Sakrabani, Ruben; Kebede, Fassil

doi:10.1016/j.trac.2020.116166

Cited by 111 publications

(64 citation statements)

References 125 publications

(140 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several different algorithms have been used for preprocessing of spectral data. Commonly used preprocessing algorithms include moving averages, binning, smoothing such as Savitzky-Golay filtering, normalization, continuum removal, derivatives, gap derivatives, multiplicative scatter and SNV computation [20][21][22][23]. Savitzky-Golay is a smoothing function which reduces noise by using a weighted sum of neighboring values, while derivatives remove additive or multiplicative effects between spectra [20].…”

Section: Referencesmentioning

confidence: 99%

“…Two types of models are used in spectral predictions; statistical-based models and machine learning-based or algorithmic models [24]. Research comparing combination of preprocessing and modeling algorithms on single spectral dataset for prediction of various soil properties are very rare and mostly studying SOC or soil clay content [21,23,25,26]. Statistical-based models are based on assumptions made by the user; while machine learning models are data driven and learn from the data set without the user assuming any parameters [27].…”

Section: Referencesmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of Optimized Preprocessing and Modeling Algorithms for Prediction of Soil Properties Using VIS-NIR Spectroscopy

Vestergaard

Vasava

Aspinall³

et al. 2021

Sensors

View full text Add to dashboard Cite

The absorbance spectra for air-dried and ground soil samples from Ontario, Canada were collected in the visible and near-infrared (VIS-NIR) region from 343 to 2200 nm. The study examined thirteen combination of six preprocessing (1st derivative, 2nd derivative, Savitzky-Golay, Gap, SNV and Detrend) method included in ‘prospectr’ R package along with four modeling approaches: partial least square regression (PLSR), cubist, random forest (RF), and extreme learning machine (ELM) for prediction of the soil organic matter (SOM). The 1st derivative + gap, 2nd derivative + gap and standard normal variance (SNV) were the best preprocessing algorithms. Thus, only these three preprocessing algorithms along with four modeling approaches were used for prediction of soil pH, electrical conductively (EC), %sand, %silt, %clay, %very coarse sand (VCS), %coarse sand (CS), %medium sand (ms) and %fine sand (fs). The results showed that OM, pH, %sand, %silt and %CS were all predicted with confidence (R2 > 0.60) and the combination of 1st derivative + gap and RF gained the best performance. A detailed comparison of the preprocessing and modeling algorithms for various soil properties in this study demonstrate that for better prediction of soil properties using VIS-NIR spectroscopy requires different preprocessing and modeling algorithms. However, in general RF and 1st derivative + gap can be labeled at the best combination of preprocessing and modelling algorithms.

show abstract

Section: Referencesmentioning

confidence: 99%

Section: Referencesmentioning

confidence: 99%

Evaluation of Optimized Preprocessing and Modeling Algorithms for Prediction of Soil Properties Using VIS-NIR Spectroscopy

Vestergaard

Vasava

Aspinall³

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Hyperspectral prediction models established using RF and ANN algorithms could estimate the SOM content in red soil plantations. In addition, PLSR, RF, SVM, and ANN are the four modeling methods shown to have excellent performance in studies on the establishment of hyperspectral prediction models [20,51,67,68], among which PLSR is the most extensively applied linear fitting method [29]. However, the relationship between the SOM content in an area and related spectral features should be more complex than a simple linear relationship.…”

Section: Comparison Of Linear and Non-linear Modeling Algorithmsmentioning

confidence: 99%

“…After reviewing the state-of-the-art literature, it is evident that exploring the most suitable band selection and selecting the appropriate combination of modeling methods are still the key problems in the field of hyperspectral prediction modeling of soil properties. ML techniques have proved to be effective in dealing with large amounts of soil spectral variables [26][27][28][29]. The techniques mentioned above have been applied to obtain prediction models of soil properties.…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral band selection and modeling of soil organic matter content in a forest using the Ranger algorithm

et al. 2021

View full text Add to dashboard Cite

Effective soil spectral band selection and modeling methods can improve modeling accuracy. To establish a hyperspectral prediction model of soil organic matter (SOM) content, this study investigated a forested Eucalyptus plantation in Huangmian Forest Farm, Guangxi, China. The Ranger and Lasso algorithms were used to screen spectral bands. Subsequently, models were established using four algorithms: partial least squares regression, random forest (RF), a support vector machine, and an artificial neural network (ANN). The optimal model was then selected. The results showed that the modeling accuracy was higher when band selection was based on the Ranger algorithm than when it was based on the Lasso algorithm. ANN modeling had the best goodness of fit, and the model established by RF had the most stable modeling results. Based on the above results, a new method is proposed in this study for band selection in the early phase of soil hyperspectral modeling. The Ranger algorithm can be applied to screen the spectral bands, and ANN or RF can then be selected to construct the prediction model based on different datasets, which is applicable to establish the prediction model of SOM content in red soil plantations. This study provides a reference for the remote sensing of soil fertility in forests of different soil types and a theoretical basis for developing portable equipment for the hyperspectral measurement of SOM content in forest habitats.

show abstract

“…the development of fast, low cost, reproducible and portable instruments available for infrared techniques (medium and near infrared) have opened new opportunities for researchers to benefit of their capabilities, especially when combined with multivariate calibrations. The latter have shown to be powerful tools to develop quantitative and qualitative models in many fields including soil 6 – 10 , food 11 , 12 pharmaceutics 13 and petroleum 14 – 18 analysis.…”

Section: Introductionmentioning

confidence: 99%

Optimizing setup of scan number in FTIR spectroscopy using the moment distance index and PLS regression: application to soil spectroscopy

Barra

Khiari

Haefele

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

Vibrational spectroscopy such as Fourier-transform infrared (FTIR), has been used successfully for soil diagnosis owing to its low cost, minimal sample preparation, non-destructive nature, and reliable results. This study aimed at optimizing one of the essential settings during the acquisition of FTIR spectra (viz. Scans number) using the standardized moment distance index (SMDI) as a metric that could trap the fine points of the curve and extract optimal spectral fingerprints of the sample. Furthermore, it can be used successfully to assess the spectra resemblance. The study revealed that beyond 50 scans the similarity of the acquisitions has been remarkably improved. Subsequently, the effect of the number of scans on the predictive ability of partial least squares regression models for the estimation of five selected soil properties (i.e., soil pH in water, soil organic carbon, total nitrogen, cation exchange capacity and Olsen phosphorus) was assessed, and the results showed a general tendency in improving the correlation coefficient (R2) as the number of scans increased from 10 to 80. In contrast, the cross-validation error RMSECV decreased with increasing scan number, reflecting an improvement of the predictive quality of the calibration models with an increasing number of scans.

show abstract

Soil spectroscopy with the use of chemometrics, machine learning and pre-processing techniques in soil diagnosis: Recent advances–A review

Cited by 111 publications

References 125 publications

Evaluation of Optimized Preprocessing and Modeling Algorithms for Prediction of Soil Properties Using VIS-NIR Spectroscopy

Evaluation of Optimized Preprocessing and Modeling Algorithms for Prediction of Soil Properties Using VIS-NIR Spectroscopy

Hyperspectral band selection and modeling of soil organic matter content in a forest using the Ranger algorithm

Optimizing setup of scan number in FTIR spectroscopy using the moment distance index and PLS regression: application to soil spectroscopy

Contact Info

Product

Resources

About