Convolutional Neural Networks as a Tool for Raman Spectral Mineral Classification Under Low Signal, Dusty Mars Conditions. Issue 10 (20th October 2022)

• Record Type: Journal Article
• Authors: Berlanga, Genesis; Williams, Quentin; Temiquel, Nathan
• Is Part Of: Earth and space science. Volume 9:Issue 10(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-10-20
• Subjects: machine learning -- Raman spectroscopy -- Mars -- mineralogy -- convolutional neural network -- CNN; Space sciences -- Periodicals; Geophysics -- Periodicals; 500.5
• DOI: 10.1029/2021EA002125
• Languages: English
• ISSNs: 2333-5084

Abstract

Abstract: NASA's Mars 2020 and ESA's ExoMars will collect Raman measurements in dusty field conditions obscuring underlying rocks. This presents a challenge for remote Raman measurements at distances where mechanical or ablative sample cleaning is not straightforward. Historically, probing broad lithostratigraphic suites has been thwarted by the need for pristine targets and high‐quality spectra. We provide a means of identifying Raman spectra of common rock‐forming silicate, carbonate, and sulfate minerals under low signal‐to‐noise‐ratios, Mars‐like conditions using a convolutional neural network (CNN). The CNN was trained on the Machine Learning Raman Open Data set data set with 500, 000+ Raman spectra of hand samples/powder mixtures (5, 000+ spectra/mineral class). Diversity in sample microtopography, orientation, and crystallinity simulated varying laser focuses and spectral quality, and no traditional spectral preprocessing such as cosmic ray or baseline removal was employed. The CNN identified low‐intensity Raman scatterers (micas and amphiboles), mixed minerals, and distinguished between mineral endmembers with +99% success. We present among the first known implementations of "big data" machine learning using varied, high‐volume Raman spectral datasets. The pattern recognition abilities of CNNs can facilitate scientist Raman spectral interpretation on Earth and autonomous rover decision‐making on planets like Mars; increasing scientific yield, correcting human classification errors, reducing the need for thorough target dust removal during evaluative measurements, and streamlining the data communications pipeline—saving time and resources. This study examines an end‐to‐end development process for creating a deep learning algorithm sensitive to varieties of Raman spectra and provides guidelines for CNN model development at the interface of Raman spectroscopy, deep learning, and planetary science. Plain Language Summary: We collected thousands of spectra from rocks and minerals that are found on Earth and Mars and trained a computer to automatically identify them using an algorithm called a convolutional neural network or CNN. This is an effective but difficult algorithm to fine tune for Raman spectra as it requires a lot of training. A CNN works like an artificial brain with firing neurons. This brain can be uploaded onto Mars rovers such as NASA's Curiosity and Perseverance so that they can automatically identify rocks and minerals on Mars without human guidance. Doing this will accelerate scientific discovery while saving crucial mission time and energy. We essentially programmed the "brains" of a rock and mineral identification "tricorder". Key Points: Identified minerals and mineral mixtures under dusty, low signal, Mars‐like conditions using Raman spectroscopy and machine learning Implemented a convolutional neural net on Machine Learning Raman Open Data set‐a dedicated Raman machine learning data set operating in a statistically significant domain General guidance given for convolutional neural network model development deployed on custom Raman spectra test sets for automated mineral identification