Sediment core analysis using artificial intelligence

Abstract

Subsurface stratigraphic modeling is crucial for a variety of environmental, societal, and economic challenges. However, the need for specific sedimentological skills in sediment core analysis may constitute a limitation. Methods based on Machine Learning and Deep Learning can play a central role in automatizing this time-consuming procedure. In this work, using a robust dataset of high-resolution digital images from continuous sediment cores of Holocene age that reflect a wide spectrum of continental to shallow-marine depositional environments, we outline a novel deep-learning-based approach to perform automatic semantic segmentation directly on core images, leveraging the power of convolutional neural networks. To optimize the interpretation process and maximize scientific value, we use six sedimentary facies associations as target classes in lieu of ineffective classification methods based uniquely on lithology. We propose an automated model that can rapidly characterize sediment cores, allowing immediate guidance for stratigraphic correlation and subsurface reconstructions.

Figure 1

Visual performance of the model on five representative images of the validation dataset. The original full-resolution digital images, the model-produced segmentation masks, and the corresponding ground truths are shown in the left, central, and right columns, respectively.

Figure 2

Visual performance of the model on five representative images of the test dataset. The original full-resolution digital images, the model produced segmentation masks, and the corresponding ground truths are shown in the left, central, and right columns, respectively. The red dots mark the images coming from the external set of sediment cores.

Figure 3

Confusion matrices for validation data (A) and test data (B). Each row of the matrices represents the instances in a ground truth class, while each column represents the class instances predicted by the model. The values were normalized with respect to the number of ground truth instances for each class. A colormap visually highlights the higher values with darker shades of blue.

Figure 4

Segmentation error and model confidence. Core images (A–F), ground truth segmentation masks (B–G), model predictions (C–H), model confidence (D–I), and segmentation error (E–L) are shown for two validation and test representative cores. The model confidence represents the prediction probability associated with the predicted class. The segmentation error is the normalized categorical cross-entropy calculated between the prediction and the ground truth.

Figure 5

(A) Examples of digital images of continuous sediment cores with associated segmentation masks (B). (C) Target classes and background colors. (D) Relative target classes abundances.