Open-Set Identification of Minerals From CRISM Hyperspectral Data

Sandeepan Dhoundiyal; Moni Shankar Dey; Shashikant Singh; Pattathal V. Arun; Guneshwar Thangjam; Alok Porwal

doi:10.1109/IGARSS53475.2024.10642309

Open-Set Identification of Minerals From CRISM Hyperspectral Data

Sandeepan Dhoundiyal, Moni Shankar Dey, Shashikant Singh, Pattathal V. Arun, Guneshwar Thangjam, Alok Porwal

Research output: Contribution to conference › Paper › peer-review

Abstract

Hyperspectral data from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) have proven instrumental in analysing the mineralogy of the Martian surface and advanced our understanding of the geological history and habitability of Mars. Recently, machine learning-based methods have been used to analyse CRISM data. However, these methods exhibit limitations such as training difficulties and the need for manual feature selection. Therefore, in the contribution we propose a novel algorithm which combines the Random Forest algorithm with Extreme Value Analysis to classify CRISM spectra under an open-set regime. The algorithm's effectiveness is demonstrated using ∼ 470 000 labelled spectra from the CRISM machine learning toolkit's mineral dataset where it returns an accuracy of 86.92 % and a Kappa(κ) of 0.85

Original language	English
Pages	6164-6167
Number of pages	4
DOIs	https://doi.org/10.1109/IGARSS53475.2024.10642309
State	Published - 1 Jan 2024
Externally published	Yes
Event	2024 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2024 - Athens, Greece Duration: 7 Jul 2024 → 12 Jul 2024

Conference

Conference	2024 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2024
Country/Territory	Greece
City	Athens
Period	7/07/24 → 12/07/24

Keywords

applied machine learning
CRISM
Hyperspectral remote sensing
mineralogical mapping
open-set classification

ASJC Scopus subject areas

Computer Science Applications
General Earth and Planetary Sciences

Access to Document

10.1109/IGARSS53475.2024.10642309

Cite this

@conference{5f9a66a312f04f13aec84528f73a28ee,

title = "Open-Set Identification of Minerals From CRISM Hyperspectral Data",

abstract = "Hyperspectral data from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) have proven instrumental in analysing the mineralogy of the Martian surface and advanced our understanding of the geological history and habitability of Mars. Recently, machine learning-based methods have been used to analyse CRISM data. However, these methods exhibit limitations such as training difficulties and the need for manual feature selection. Therefore, in the contribution we propose a novel algorithm which combines the Random Forest algorithm with Extreme Value Analysis to classify CRISM spectra under an open-set regime. The algorithm's effectiveness is demonstrated using ∼ 470 000 labelled spectra from the CRISM machine learning toolkit's mineral dataset where it returns an accuracy of 86.92 % and a Kappa(κ) of 0.85",

keywords = "applied machine learning, CRISM, Hyperspectral remote sensing, mineralogical mapping, open-set classification",

author = "Sandeepan Dhoundiyal and Dey, {Moni Shankar} and Shashikant Singh and Arun, {Pattathal V.} and Guneshwar Thangjam and Alok Porwal",

note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 2024 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2024 ; Conference date: 07-07-2024 Through 12-07-2024",

year = "2024",

month = jan,

day = "1",

doi = "10.1109/IGARSS53475.2024.10642309",

language = "English",

pages = "6164--6167",

}

TY - CONF

T1 - Open-Set Identification of Minerals From CRISM Hyperspectral Data

AU - Dhoundiyal, Sandeepan

AU - Dey, Moni Shankar

AU - Singh, Shashikant

AU - Arun, Pattathal V.

AU - Thangjam, Guneshwar

AU - Porwal, Alok

PY - 2024/1/1

Y1 - 2024/1/1

N2 - Hyperspectral data from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) have proven instrumental in analysing the mineralogy of the Martian surface and advanced our understanding of the geological history and habitability of Mars. Recently, machine learning-based methods have been used to analyse CRISM data. However, these methods exhibit limitations such as training difficulties and the need for manual feature selection. Therefore, in the contribution we propose a novel algorithm which combines the Random Forest algorithm with Extreme Value Analysis to classify CRISM spectra under an open-set regime. The algorithm's effectiveness is demonstrated using ∼ 470 000 labelled spectra from the CRISM machine learning toolkit's mineral dataset where it returns an accuracy of 86.92 % and a Kappa(κ) of 0.85

AB - Hyperspectral data from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) have proven instrumental in analysing the mineralogy of the Martian surface and advanced our understanding of the geological history and habitability of Mars. Recently, machine learning-based methods have been used to analyse CRISM data. However, these methods exhibit limitations such as training difficulties and the need for manual feature selection. Therefore, in the contribution we propose a novel algorithm which combines the Random Forest algorithm with Extreme Value Analysis to classify CRISM spectra under an open-set regime. The algorithm's effectiveness is demonstrated using ∼ 470 000 labelled spectra from the CRISM machine learning toolkit's mineral dataset where it returns an accuracy of 86.92 % and a Kappa(κ) of 0.85

KW - applied machine learning

KW - CRISM

KW - Hyperspectral remote sensing

KW - mineralogical mapping

KW - open-set classification

UR - http://www.scopus.com/inward/record.url?scp=85208475795&partnerID=8YFLogxK

U2 - 10.1109/IGARSS53475.2024.10642309

DO - 10.1109/IGARSS53475.2024.10642309

M3 - Paper

AN - SCOPUS:85208475795

SP - 6164

EP - 6167

T2 - 2024 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2024

Y2 - 7 July 2024 through 12 July 2024

ER -

Open-Set Identification of Minerals From CRISM Hyperspectral Data

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this