Determination of soil pH from Vis-NIR spectroscopy by extreme learning machine and variable selection: A case study in lime concretion black soil

Wang, Liusan; Wang, Rujing

doi:10.1016/j.saa.2022.121707

Cited by 26 publications

(7 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The flow has a significant impact on soil erosion, lifting up the material and destroying the morphology of the soil surface [ 15 , 22 ]. When the flow is large enough, the material carried by the flow abrades the surface, and a larger flow always leads to stronger soil erosion [ 74 , 75 ]. As can be seen, the flow sediment content (CS) reflects the soil erosion efficiency [ 74 ].…”

Section: Discussionmentioning

confidence: 99%

“…When the flow is large enough, the material carried by the flow abrades the surface, and a larger flow always leads to stronger soil erosion [ 74 , 75 ]. As can be seen, the flow sediment content (CS) reflects the soil erosion efficiency [ 74 ]. This study showed that under short-term heavy rainfall conditions, the sediment content had an extremely significant negative correlation with τ ( r = −0.863 **) and DW-f ( r = −0.863 **) and a significant negative correlation with Re ( r = −0.735 *) but showed no significant correlation with other hydraulic characteristics.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

The Influence of Short-Term Heavy Rainfall on Hydraulic Characteristics and Rill Formation in the Yuanmou Dry-Hot Valley

Luo

Ma²,

Wang³

et al. 2022

IJERPH

View full text Add to dashboard Cite

Rill erosion is one of the major environmental problems in the world; it is an important factor with regard to land degradation and has a serious impact on production and daily life in the region. The widely distributed Yuanmou group stratum promotes the development of rill erosion, whereby the strong time-concentrated rainfall and the alternating arid-humid climate prepare the ground for the development of rills in soils. Therefore, a study of the processes of slope rill erosion was carried out, and a gravel-soil slope in the Yuanmou dry-hot valley was chosen to simulate short-term heavy rainfall (25 mm/h) (No. 1 plot) and moderate rainfall (15 mm/h) (No. 2 plot), to study the erosion processes of soil and the dynamic characteristics of runoff involved in erosion. The study results showed that the width of runoff was significantly different between the two plots, while the depth of runoff was not significantly different. During the rill formation process, the width of the two plots first decreased and then increased with increasing washout duration, while its depth did not change significantly. Flow was the key factor in determining the hydraulic characteristics of runoff, and it had a significant or extremely significant positive correlation with hydraulic characteristics parameters, except in the case of Fr (Froude number) (r = 0.039). The total sediment content (CS) of plot No. 1 (0.158 g/cm3) was significantly different from that of plot No. 2 (0.153 g/cm3), and both CSs in the two plots decreased with increasing washout duration. The CS had an extremely significant negative correlation with τ (runoff shear force) (r = −0.863 **) and DW-f (Darcy-Weisbach drag coefficient) (r = −0.863 **) and a significant negative correlation with Re (Reynolds number) (r = −0.735 *) in the short-term heavy rainfall experiment, while the CS had a significant positive correlation with V (velocity) (r = 0.814 *), R (hydraulic radius) (r = 0.811 *) and P (unit stream power) (r = 0.811 *) in the moderate rainfall experiment. The results of this study will help guide further examination of the processes involved in the dynamic mechanisms of rill erosion on slopes under short-term heavy rainfall conditions.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

The Influence of Short-Term Heavy Rainfall on Hydraulic Characteristics and Rill Formation in the Yuanmou Dry-Hot Valley

Luo

Ma²,

Wang³

et al. 2022

IJERPH

View full text Add to dashboard Cite

show abstract

“…The process of calculating the coefficient stability values of noisy data is also skipped by MCUVE by eliminating features with low coefficient stability values. To predict pH in lime concretion black soil, the study in [53] used CWT to preprocess the soil spectra, followed by ELM combined with four spectral variable selection methods, GA, SPA, MCUVE, and CARS, and the full spectrum. According to the results of the experiment, the MCUVE feature selection algorithm had the lowest residual prediction deviation.…”

Section: Monte Carlo Uninformative Variable Elimination (Mcuve)mentioning

confidence: 99%

A Review of Machine Learning for Near-Infrared Spectroscopy

Zhang

Kasun

Wang

et al. 2022

Sensors

View full text Add to dashboard Cite

The analysis of infrared spectroscopy of substances is a non-invasive measurement technique that can be used in analytics. Although the main objective of this study is to provide a review of machine learning (ML) algorithms that have been reported for analyzing near-infrared (NIR) spectroscopy from traditional machine learning methods to deep network architectures, we also provide different NIR measurement modes, instruments, signal preprocessing methods, etc. Firstly, four different measurement modes available in NIR are reviewed, different types of NIR instruments are compared, and a summary of NIR data analysis methods is provided. Secondly, the public NIR spectroscopy datasets are briefly discussed, with links provided. Thirdly, the widely used data preprocessing and feature selection algorithms that have been reported for NIR spectroscopy are presented. Then, the majority of the traditional machine learning methods and deep network architectures that are commonly employed are covered. Finally, we conclude that developing the integration of a variety of machine learning algorithms in an efficient and lightweight manner is a significant future research direction.

show abstract

“…Comprehensive spectral libraries are being generated in different countries to develop robust chemometric models (Bellinaso et al, 2010; Shepherd et al, 2022; Stevens et al, 2013; Viscarra Rossel et al, 2016). Large variability in spectral database (Coblinski et al, 2020), robust chemometric models (Ludwig et al, 2019; Wang & Wang, 2022) and different pre‐processing approaches (Dotto et al, 2018) have improved the performance of the DRS approach. Improvements in the performance of chemometric models have also been reported by combining both VNIR and the mid‐infrared (MIR) spectra (Viscarra Rossel et al, 2006) and also by combining VNIR spectra of different aggregate size fractions (Vasava et al, 2019; Vasava & Das, 2022).…”

Section: Introductionmentioning

confidence: 99%

“…Comprehensive spectral libraries are being generated in different countries to develop robust chemometric models (Bellinaso et al, 2010;Stevens et al, 2013;Viscarra Rossel et al, 2016;Shepherd et al, 2022). Large variability in spectral database (Coblinski et al, 2020), robust chemometric models (Ludwig et al, 2019;Wang and Wang, 2022) and different pre-processing approaches (Dotto et al, 2018) have improved the performance of the DRS approach.…”

Section: Introductionmentioning

confidence: 99%

Diffuse reflectance spectroscopy (DRS) for rapid soil testing and soil quality assessment in smallholder farms

Majeed

Garg

Venkataradha

et al. 2023

European J Soil Science

View full text Add to dashboard Cite

Rapid soil testing and soil quality assessment are essential to address soil degradation and low farm incomes in smallholder farms. With the objective of testing diffuse reflectance spectroscopy (DRS) to rapidly assess soil chemical properties, nutrient content and a soil quality index (SQI), samples of surface soil were collected from 1113 smallholder farms in seven districts in Bundelkhand region of Uttar Pradesh, India. A minimum dataset (MDS) approach was followed to estimate SQI using the three chemical parameters of soil pH, electrical conductivity (EC) and soil organic carbon (SOC), and 11 different soil nutrients. Principal component and correlation analyses showed that soil pH, SOC content and three available nutrients À copper (Cu), iron (Fe) and sulphur (S) À may constitute the MDS. Estimated SQI values showed strong positive correlation with crop yields. Results of chemometric modelling showed that the DRS approach could yield the coefficient of determination (R 2 ) values in the validation datasets ranging from 0.79 to 0.94 for exchangeable calcium (Ca) followed by 0.67-0.88 for exchangeable potassium (K), 0.52-0.86 for SOC and 0.53-0.81 for available boron (B) content. Except in one district, the DRS approach could be used to estimate SQI values with R 2 values in the range of 0.63-0.81; an R 2 value of 0.71 was obtained in the pooled dataset. We also estimated the three-tier soil test crop response (STCR) ratings to compare DRS and wet chemistry soil testing approaches. Similar STCR ratings were obtained for both these approaches in more than 86% of the samples. Parameters for which both the methods yielded similar ratings in more than 80% of the samples were EC (>98%), pH and exchangeable Ca (>81%) and available B (>89%). With similar ratings, these results suggest that the DRS approach may safely be used for farmers' fields, replacing the traditional wet analysis approach of soil testing.

show abstract

Determination of soil pH from Vis-NIR spectroscopy by extreme learning machine and variable selection: A case study in lime concretion black soil

Cited by 26 publications

References 56 publications

The Influence of Short-Term Heavy Rainfall on Hydraulic Characteristics and Rill Formation in the Yuanmou Dry-Hot Valley

The Influence of Short-Term Heavy Rainfall on Hydraulic Characteristics and Rill Formation in the Yuanmou Dry-Hot Valley

A Review of Machine Learning for Near-Infrared Spectroscopy

Diffuse reflectance spectroscopy (DRS) for rapid soil testing and soil quality assessment in smallholder farms

Contact Info

Product

Resources

About