Hard rock cores recovered during International Ocean Discovery Program South Atlantic Tran- sect (SAT) Expeditions 390C, 395E, 390, and 393 were digitally imaged using a DMT CoreScan3 line scanner aboard the R/V JOIDES Resolution during Expeditions 390 and 393. Methods out- lined here involve modification of the CoreScan3 to scan >1 m long sections of hard rock at a 40 pixel/mm resolution, four times higher resolution than previously achieved during an ocean drilling campaign. This method is combined with a dedicated Python package (StitchIT) written to compile multiple 20 cm frames into high-resolution core section images. Accompanying this report are all compiled high-resolution section images from the SAT, both as high-resolution individual images of each section and as moderate-resolution visual core description–style images. The methods outline best practices for digitally imaging external core surfaces, and the data set is the first to be generated specifically for emerging machine learning and computer vision applications.
Data report: high-resolution digital imaging of whole-round hard rocks collected during IODP Expeditions 390C, 395E, 390, and 393, South Atlantic Transect, using a DMT CoreScan3
Expedition 390/393 ScientistsMembro del Collaboration Group
2024-01-01
Abstract
Hard rock cores recovered during International Ocean Discovery Program South Atlantic Tran- sect (SAT) Expeditions 390C, 395E, 390, and 393 were digitally imaged using a DMT CoreScan3 line scanner aboard the R/V JOIDES Resolution during Expeditions 390 and 393. Methods out- lined here involve modification of the CoreScan3 to scan >1 m long sections of hard rock at a 40 pixel/mm resolution, four times higher resolution than previously achieved during an ocean drilling campaign. This method is combined with a dedicated Python package (StitchIT) written to compile multiple 20 cm frames into high-resolution core section images. Accompanying this report are all compiled high-resolution section images from the SAT, both as high-resolution individual images of each section and as moderate-resolution visual core description–style images. The methods outline best practices for digitally imaging external core surfaces, and the data set is the first to be generated specifically for emerging machine learning and computer vision applications.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
