Deep Learning Associated with Laser-Induced Breakdown Spectroscopy (LIBS) for the Prediction of Lead in Soil

Abstract

In this study, a method based on laser-induced breakdown spectroscopy (LIBS) was developed to detect soil contaminated with Pb. Different levels of Pb were added to soil samples in which tobacco was planted over a period of two to four weeks. Principal component analysis and deep learning with a deep belief network (DBN) were implemented to classify the LIBS data. The robustness of the method was verified through a comparison with the results of a support vector machine and partial least squares discriminant analysis. A confusion matrix of the different algorithms shows that the DBN achieved satisfactory classification performance on all samples of contaminated soil. In terms of classification, the proposed method performed better on samples contaminated for four weeks than on those contaminated for two weeks. The results show that LIBS can be used with deep learning for the detection of heavy metals in soil.

Discrimination of Grape Seeds Using Laser-Induced Breakdown Spectroscopy in Combination with Region Selection and Supervised Classification Methods

“…In some previous studies in spectroscopic tasks, deep learning methods routinely outperformed other traditional machine learning methods [33,34]. In terms of deep learning in LIBS applications, Zhao et al attempted to identify different concentrations of lead in soil where tobacco was grown to absorb the contamination [15]. The deep belief network (DBN) classifier performed better than the SVM and partial least squares-discriminant analysis (PLS-DA) models for samples contaminated for two and four weeks.…”

“…On the other hand, deep learning, which represents a wide class of machine learning methods mostly based on artificial neural networks, has become the most popular topic in the artificial intelligence field. For LIBS data analysis to date, attention has mostly been paid to the implementation of traditional machine learning algorithms [7,11,12], while only a few studies have searched for a way to interpret LIBS data by deep learning approach [15]. To our knowledge, no attempt has been made to evaluate the feasibility of deep learning for discriminating grape seeds.…”

Application of Scikit and Keras Libraries for the Classification of Iron Ore Data Acquired by Laser-Induced Breakdown Spectroscopy (LIBS)

“…As the most basic structure in the deep learning field, DNN classifies samples by setting activation function, loss function, etc. [34]. Theoretically, DNN can solve any classification problem.…”

“…At different levels of Pb, the results show that deep learning can handle LIBS data and highlight the importance of using samples. The total accuracy values are 98.47% (training set) and 90.625% (test set) for samples [34].…”

Lithium Isotope Measurement Using Laser-Induced Breakdown Spectroscopy and Chemometrics

“…Chemometric techniques are often applied to LIBS spectra to quantitatively determine analyte concentrations. Some recent examples include the determination of pH and multiple analyte concentrations in topsoil, 29–31 analyte concentrations in pharmaceutical quality control, 28 fly ash in boiler efficiency measurement, 32 and even plastic classification in recycling management. 33…”