Quantification of Hydrocarbon Abundance in Soils Using Deep Learning with Dropout and Hyperspectral Data

Ahmed, Asmau M.; Durán, Olga; Zweiri, Yahya; Smith, Mike

doi:10.3390/rs11161938

Cited by 6 publications

(5 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dropout forces the network to be less sensitive to the specific weight of neurons. According to Ahmed et al (2019), Dropout has promising generalization and using it, the network is less likely to become overfitted.…”

Section: Fnn Spectral Modellingmentioning

confidence: 99%

National spectral data and learning algorithms for potentially toxic elements modelling in forest soil horizons

Gholizadeh¹,

Saberioon²,

Dor³

et al. 2020

Preprint

View full text Add to dashboard Cite

Forest ecosystems are among the main parts of the biosphere; however, they have been endangered from the significant elevation and harmful effects of air and soil pollutants, including potentially toxic elements (PTEs). The concentration of PTEs in forest soils varies not only laterally but also vertically with depth. Forest surface organic horizons are of particular interest in forest ecosystem monitoring due to their role as stable adsorbents of the deposited atmospheric substances. Therefore, the main purpose of this study was to conduct rapid examinations of forest soils PTEs (Cr, Cu, Pb, Zn, and Al), testing the capability of VIS--NIR spectroscopy coupled with machine learning (ML) techniques (partial least square regression (PLSR), support vector machine regression (SVMR), and random forest (RF)) and fully connected neural network (FNN), a deep learning (DL) approach, in forest organic horizons. One-thousand-and-eighty forested sites across the Czech Republic at two soil layers, defining the fragmented (F) and humus (H) organic horizons, were investigated (total 2160 samples). PTEs as well as total Fe and SOC, as auxiliary data, were conventionally and spectrally determined and modelled in the combined organic horizons (F + H) and in each individual horizon using the ML and DL algorithms. Results indicated that the concentration of all PTEs was higher in the horizon H compared to the F horizon. Although the spectral reflectance of samples tended to decrease with increased PTEs concentration. Strongly significant positive correlations between all PTEs and total Fe in all horizons were obtained, which were higher in the H and F + H horizons than the F horizon. The highest correlations of PTEs with the spectra were at 460--590~nm, which is mostly linked to the presence of Fe-oxide. These results show the importance of Fe for spectral prediction of PTEs. Cr and Al were the most accurately predicted elements, regardless of the applied learning technique. SVMR provided the best results in assessing the H horizon (e.g., R\(^2\) = 0.88 and root mean square error (RMSE) = 3.01~mg/kg, and R\(^2\) = 0.82 and RMSE = 1682.25~mg/kg for Cr and Al, respectively); however, FNN predicted the combined F + H horizons the best (R\(^2\) = 0.89 and RMSE = 2.95~mg/kg, and R\(^2\) = 0.86 and RMSE = 1593.64~mg/kg for Cr and Al, respectively) due to the larger number of samples. In the F horizon, almost no parameters were predicted adequately. This study shows that given the availability of larger sample sizes, FNN can be a more promising technique compared to ML methods for assessment of Cr and Al concentration based on national spectral data in the forests of the Czech Republic.

show abstract

Section: Fnn Spectral Modellingmentioning

confidence: 99%

National spectral data and learning algorithms for potentially toxic elements modelling in forest soil horizons

Gholizadeh¹,

Saberioon²,

Dor³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…On the other hand, Visible-Near Infrared (Vis-NIR) spectroscopy is proven to be a portable, quick, and non-invasive method for hydrocarbon assessment in soil with good accuracy [5]. In the Vis-NIR spectroscopy range (800-2500 nm), hydrocarbon spectra originate mainly from combinations or overtones of C-H stretching modes of saturated CH2 and terminal CH3 or aromatic C-H functional groups resulting in absorption at 1200, 1725, and 2310 nm [5,6]. However, spectroscopy alone fails to give information on spatial distribution of the spectra obtained which limits its application [7].…”

Section: Introductionmentioning

confidence: 99%

“…Due to the direct relation between PHCs and their reflectance spectra, the spectra can be used to make predictions on the levels of PHC contamination in soil mediums quantitatively and qualitatively by analyzing it [5,6]. Several different methods have been used to analyze spectral data to assess petroleum hydrocarbons in soils.…”

Section: Introductionmentioning

confidence: 99%

“…However, factors affecting hyperspectral remote sensing such as spectral variability due to varying image conditions and mixed pixel spectra due to contributions from different materials introduce complexities that require nonlinear analysis methods [11]. Nonlinear methods for spectral data analysis include neural networks and kernel-based transformations [6,12]. This paper's objective is to build predictive models to estimate the amount of petroleum hydrocarbon contaminant in contaminated soils using hyperspectral images while making use of spatial and spectral features through deep neural networks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Quantification of Hydrocarbon Contamination in Soil Using Hyperspectral Data and Deep Learning

Ayass¹,

Mustapha²,

Salam³

2023

World Congress on Civil, Structural, and Environmental Engineering

View full text Add to dashboard Cite

Petroleum and its products undergo large scale production, transportation, and storage which make them prone to spills and leakages into the environment. Petroleum contamination in terrestrial environments, particularly soil bodies, is common and holds major consequences on food and crops, microbial communities, the atmosphere, the water sphere, public health and safety, and the soil itself, and therefore, requires immediate detection and assessment in case contamination is present. In this paper, advanced machine learning methods are used to predict petroleum hydrocarbon contamination in soil using hyperspectral image data. Hyperspectral imaging combines imaging and spectroscopy and can detect petroleum hydrocarbons using the characteristic absorption features in hydrocarbon reflectance spectra. Laboratory prepared soil samples are contaminated with crude oil and scanned with a hyperspectral camera in a laboratory setup. The data collected is used to train deep learning models to predict, quantitatively, the amount of petroleum present in soil samples. To make predictions, a first model is built to use spectral data from a single pixel while a second model is built to use spectral and spatial data by using two adjacent pixel spectra as input. The results show good performance for both models, with the twopixel model achieving a better mean square error of 0.48 on a testing dataset compared to the mean square error of 0.628 for the single pixel model on the same testing data. Therefore, hyperspectral imaging contains valid spectral and spatial information that are beneficial for assessing petroleum contamination in soil with good accuracy.

show abstract

“…The traditional methods for determining the Ni content in soil requires field sampling followed by laboratory analysis, but it is time-consuming, costly, and inefficient [9,10]. Hyperspectral remote sensing technology has been applied to predict the physical, chemical, and biological properties of soil, due to the advantages of rapid, accurate, nondestructive, low-cost, and dynamic monitoring over a large area [11][12][13], including soil moisture [14,15], hydrocarbon content [16], nitrogen content [17], organic matter content [18], electrical conductivity [19], and salt content [20,21]. In recent years, hyperspectral remote sensing technology has shown good results in the prediction of Ni content in soil [22,23].…”

Section: Introductionmentioning

confidence: 99%

Application of a Hyperspectral Remote Sensing Model for the Inversion of Nickel Content in Urban Soil

Zhong,

Eziz,

Sawut

et al. 2023

Sustainability

View full text Add to dashboard Cite

Hyperspectral remote sensing technology can provide a rapid and nondestructive method for soil nickel (Ni) content detection. In order to select a high-effective method for estimating the soil Ni content using a hyperspectral remote sensing technique, 88 soil samples were collected in Urumqi, northwest China, to obtain Ni contents and related hyperspectral data. At first, 12 spectral transformations were used for the original spectral data. Then, Pearson’s correlation coefficient analysis (PCC) and the CARS method were used for selecting important wavelengths. Finally, partial least squares regression (PLSR), random forest regression (RFR) and support vector machine regression (SVMR) models were used to establish the hyperspectral inversion models of the Ni content in the soil using the important wavelengths. The coefficient of determination (R2), root mean square error (RMSE), mean absolute error (MAE), and residual prediction deviation (RPD) were selected to evaluate the inversion effects of the models. The results indicated that using the PCC and CARS method for the original and transformed wavebands can effectively improve the correlations between the spectral data and Ni content of the soil in the study area. The random forest regression model, based on the first-order differentiation of the reciprocal (RTFD–RFR), was more stable and had the best inversion effects, with the highest predictive ability (R2 = 0.866, RMSE = 1.321, MAE = 0.986, RPD = 2.210) for determining the Ni content in the soil. The RTFD–RFR methods can be used as a means of the inversion of the Ni content in urban soil. The results of the study can provide a technical support for the hyperspectral estimation of the Ni content of urban soil.

show abstract

Quantification of Hydrocarbon Abundance in Soils Using Deep Learning with Dropout and Hyperspectral Data

Cited by 6 publications

References 63 publications

National spectral data and learning algorithms for potentially toxic elements modelling in forest soil horizons

National spectral data and learning algorithms for potentially toxic elements modelling in forest soil horizons

Quantification of Hydrocarbon Contamination in Soil Using Hyperspectral Data and Deep Learning

Application of a Hyperspectral Remote Sensing Model for the Inversion of Nickel Content in Urban Soil

Contact Info

Product

Resources

About