Granite Extraction Based on the SDGSAT-1 Satellite Thermal Infrared Spectrometer Imagery

Earth observation by remote sensing plays a crucial role in granite extraction, and many current studies use thermal infrared data from sensors such as ASTER. The challenge lies in the low spatial resolution of these satellites, hindering precise rock type identification. A breakthrough emerges with the Thermal Infrared Spectrometer (TIS) on the Sustainable Development Science Satellite 1 (SDGSAT-1) launched by the Chinese Academy of Sciences. With an exceptional 30 m spatial resolution, SDGSAT-1 TIS opens avenues for accurate granite extraction using remote sensing. This study, exemplified in Xinjiang's Karamay region, introduces the BR-ISauvola method, leveraging SDGSAT-1 TIS data. The approach combines band ratio with adaptive k-value selection using local grayscale statistical features for Sauvola thresholding. Focused on large-scale granite extraction, results show F1 scores above 70% for Otsu, Sauvola, and BR-ISauvola. Notably, BR-ISauvola achieves the highest accuracy at 82.11%, surpassing Otsu and Sauvola by 9.62% and 0.34%, respectively. This underscores the potential of SDGSAT-1 TIS data as a valuable resource for granite extraction. The proposed method efficiently utilizes spectral information, presenting a novel approach for rapid granite extraction using remote sensing TIS imagery, even in scenarios with low spectral resolution and a single data source.

Medienart:

E-Artikel

Erscheinungsjahr:

2024

Erschienen:

2024

Enthalten in:

Zur Gesamtaufnahme - volume:24

Enthalten in:

Sensors (Basel, Switzerland) - 24(2024), 6 vom: 08. März

Sprache:

Englisch

Beteiligte Personen:

Yuan, Boqi [VerfasserIn]
Wang, Qinjun [VerfasserIn]
Yang, Jingyi [VerfasserIn]
Xu, Wentao [VerfasserIn]
He, Chaokang [VerfasserIn]

Links:

Volltext

Themen:

Band ratio
Granite extraction
Journal Article
Karamay
SDGSAT-1 TIS
Sauvola threshold

Anmerkungen:

Date Revised 30.03.2024

published: Electronic

Citation Status PubMed-not-MEDLINE

doi:

10.3390/s24061750

funding:

Förderinstitution / Projekttitel:

PPN (Katalog-ID):

NLM370335503