Automatic preprocessing of laser-induced breakdown spectra using partial least squares regression and feed-forward artificial neural network: Applications to Earth and Mars data

Ewusi-Annan, Ebo; Delapp, D.; Wiens, R. C.; Melikechi, Noureddine

doi:10.1016/j.sab.2020.105930

Cited by 24 publications

(8 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the influence of the preprocessing on the results could be studied by using raw data. The combination with machine learning models for preprocessing (Castorena et al, 2021;Ewusi-Annan et al, 2020) is also conceivable. Another aspect, which was not discussed in detail here is the investigation of the rejected observation points.…”

Section: Discussionmentioning

confidence: 99%

“…A study using the data from 376 of these standards to train a convolutional neural network (CNN) with different activation functions reveals promising results for the prediction of major oxide compositions (Cao et al, 2020). Machine learning algorithms are not only promising for regression and classification purposes, they can also support the preprocessing of ChemCam data (Castorena et al, 2021;Ewusi-Annan et al, 2020) and enable autonomous ChemCam target selection with the Autonomous Exploration for Gathering Increased Science (AEGIS) targeting system (Francis et al, 2017).…”

Section: 1029/2021ea001903mentioning

confidence: 99%

See 1 more Smart Citation

Clustering Supported Classification of ChemCam Data From Gale Crater, Mars

Rammelkamp

Gasnault

Forni

et al. 2021

Earth and Space Science

Self Cite

View full text Add to dashboard Cite

The Mars Science Laboratory (MSL) rover Curiosity landed successfully in Gale crater on Mars on August 6, 2012. One of the primary mission goals of MSL is to determine whether Mars was ever habitable, for which the Gale crater and in particular its central mountain Aeolis Mons (informally known as Mt. Sharp) was chosen as the science target (Grotzinger et al., 2012). From orbit, it was recognized that the lower layered sedimentary strata of Mt. Sharp contains a record of environmental changes including diverse aqueous conditions over time (Milliken et al., 2010). Curiosity landed at the Bradbury site, northwest of Mt. Sharp, near the base of the alluvial fan of Peace Vallis, and explored the rocks of Aeolis Palus during the first Mars year of the mission. On sol 753, Curiosity reached the geologic transition to the Mt. Sharp group and has continued to climb the mound since then. The ancient habitability could be confirmed at Yellowknife Bay, a location in Aeolis Palus, due to several detections of early diagenesis products, which together suggest the existence of an aqueous environment, implying a higher potential for habitable conditions. (Grotzinger et al., 2014(Grotzinger et al., , 2015.Several scientific instruments belong to Curiosity's payload for the characterization of Gale crater's geological evolution and one of them is the "Chemistry and Camera" (ChemCam) instrument (Maurice et al., 2012;Wiens et al., 2012). ChemCam has the capability to perform rapid multi-elemental analyses of rocks and soils from a distance with the laser-induced breakdown spectroscopy (LIBS) technique. Due to its regular use-almost every sol-ChemCam has collected a very large and unique LIBS data set containing more than 800,000 single shot LIBS spectra. Such a data set is well suited for the application of machine learning algorithms in order to identify similar targets and predict classes of new measured targets. From

show abstract

Section: Discussionmentioning

confidence: 99%

Section: 1029/2021ea001903mentioning

confidence: 99%

Clustering Supported Classification of ChemCam Data From Gale Crater, Mars

Rammelkamp

Gasnault

Forni

et al. 2021

Earth and Space Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…3.1.3 Novel LIBS approaches. Automated or semiautomated preprocessing of LIBS data, from samples of similar composition, was developed by Ewusi-Annan et al, 121 using machine learning tools. For this purpose, PLS regression and articial neural networks were applied on two LIBS datasets, obtaining relatively high accuracy on the prediction of preprocessed spectra of geological samples analysed by a laboratory model of ChemCam and by ChemCam as it interrogates Martian targets, respectively.…”

Section: Laser Induced Breakdown Spectroscopy (Libs)mentioning

confidence: 99%

Atomic spectrometry update: review of advances in atomic spectrometry and related techniques

Evans

Pisonero

Smith

et al. 2021

J. Anal. At. Spectrom.

View full text Add to dashboard Cite

show abstract

“…Thus, Yaroshchyk's MF was the best overall choice. Ewusi-Annan et al 14 proposed two automatic pre-processing algorithms based on the partial least squares (PLS) and feed-forward artificial neural network (FFANN), respectively. Yet, the limitation of those two algorithms requires a training data set.…”

Section: Introductionmentioning

confidence: 99%