Multi-Sensor Spectral Imaging of Geological Samples: A Data Fusion Approach Using Spatio-Spectral Feature Extraction

Abstract

Rapid, efficient and reproducible drillcore logging is fundamental in mineral exploration. Drillcore mapping has evolved rapidly in the recent decade, especially with the advances in hyperspectral spectral imaging. A wide range of imaging sensors is now available, providing rapidly increasing spectral as well as spatial resolution and coverage. However, the fusion of data acquired with multiple sensors is challenging and usually not conducted operationally. We propose an innovative solution based on the recent developments made in machine learning to integrate such multi-sensor datasets. Image feature extraction using orthogonal total variation component analysis enables a strong reduction in dimensionality and memory size of each input dataset, while maintaining the majority of its spatial and spectral information. This is in particular advantageous for sensors with very high spatial and/or spectral resolution, which are otherwise difficult to jointly process due to their large data memory requirements during classification. The extracted features are not only bound to absorption features but recognize specific and relevant spatial or spectral patterns. We exemplify the workflow with data acquired with five commercially available hyperspectral sensors and a pair of RGB cameras. The robust and efficient spectral-spatial procedure is evaluated on a representative set of geological samples. We validate the process with independent and detailed mineralogical and spectral data. The suggested workflow provides a versatile solution for the integration of multi-source hyperspectral data in a diversity of geological applications. In this study, we show a straight-forward integration of visible/near-infrared (VNIR), short-wave infrared (SWIR) and long-wave infrared (LWIR) data for sensors with highly different spatial and spectral resolution that greatly improves drillcore mapping.

Integrated Geological and Geophysical Mapping of a Carbonatite-Hosting Outcrop in Siilinjärvi, Finland, Using Unmanned Aerial Systems

“…We perform data fusion on a "noisy" outcrop to reduce ambiguity of interpretation while increasing detection confidence and accuracy of classifications [39]. The feasibility of such a fusion approach was laid out for different lithologies at laboratory scale where multi-source hyperspectral and photogrammetric techniques were combined [40]. We apply spatially constrained feature extraction on the UAS-based optical imagery for a consistent classification as part of our multi-sensor data approach to enhance image classification results.…”

Drill-Core Hyperspectral and Geochemical Data Integration in a Superpixel-Based Machine Learning Framework

“…This integration allows a more comprehensive and complete analysis including not only the elements that influence the diagnostic spectrum absorptions in the VNIR-SWIR but also in the LWIR. There are several studies in the literature confirming the advantages of such integration [45], [47]- [51]. In this paper, we suggest to use the composite kernel Support Vector Machine (ckSVM) algorithm [52], [53] to balance the spectral information contained in the hyperspectral data covering the VNIR-SWIR and LWIR regions of the electromagnetic spectrum.…”

Multiple Optical Sensor Fusion for Mineral Mapping of Core Samples

“…Sensor-specific variations in spatial resolution, however, provide not only a challenge for data alignment, but also result in different mineral mixtures to be represented by one pixel. In a previous publication [ 17 ], we approached this challenge by classifying meaningful mixed mineralogical domains instead of using a conventional, direct mineral mapping approach. This allowed us to include spectrally inactive minerals into the classification process and to map domains of interest which are not characterized by one specific mineral, but by an indicative mineral mixture.…”

“…Sparse and smooth low-rank analysis (SSLRA) [ 19 ] and orthogonal total variation component analysis (OTVCA) [ 20 ] were recently proposed in the literature (see [ 21 ] for an overview) and regarded as the state-of-the-art unsupervised feature extraction approaches. OTVCA [ 20 ] was used in [ 17 ] for the classification of core samples using a fusion of multisensor images. From here on, the proposed method in [ 17 ] is referred to as OTVCA_Fus.…”