Sedimentary structure discrimination with hyperspectral imaging in sediment cores

The LDB17_P11Ax (IGSN: TOAE0000000243); Datation Age 1040 +/- 30 to 2017 CE by core correlation, 14C, lamina counting) core from the Bourget Lake (France) was analyzed in 2018 by hyperspectral imaging. We studied the potential of hyperspectral sensor to image a sediment cores and created machine learning models. The hyperspectral images were acquired in order to develop quantitative (estimating particle size and loss on ignition) and qualitative (detection of instantaneous events or lamina) methods. All these methods allow to reconstruct the past environment and climate at high resolution (pixel size: 50-250 microns) and without destroying the sample for archiving for future analysis. These images have been valorized in publications for the detection of instantaneous events with hyperspectral and combined with XRF data, for the combination of the two images into a composite image. image (.hdr, .dat, .jpg