The scope of work was focused on automated core photo classification for drillholes in the Neves-Corvo mine. The aim of this project was to enhance the identification of shear zones (characterized by foliation and veining) and brittle faults (marked by broken core, rubble, and gouge).
SRK obtained the following core photo classifications: Pseudo-RQD and break frequency based on the detection of breaks and rubble zones were generated at 1 m intervals. It is important to note that both natural (fracture) and mechanical breaks were identified; quartz or carbonate veins are logged and summarized as a vein intensity table with 1 m intervals; a custom gouge logging model trained on Neves-Corvo mine core images; and qualitative designation of the degree of broken core with the Core Damage Index (CDI) model. The CDI model was developed to support geotechnical investigations by providing a measure of the degree of core damage at high resolution intervals.
SRK’s CDI model classified known fault zones as “Partially Broken Core”, "Heavily Broken Core" and "Disintegrated Core," while also identifying additional potential fault zones across the dataset. Further validation came from the pseudo-RQD and break frequency data, which aligned with the CDI results, although the CDI model provided higher-resolution insights.
SRK's vein density model revealed zones of veining that extend laterally across multiple drillholes, including an extensive vein-rich interval associated with shearing and occurring on the hanging wall of the sulphide mineralization. The gouge classification model identified many thin intervals of gouge that were composited into “gouge zones” when multiple gouge horizons were stacked in close proximity within a drillhole. Most of these gouge zones correlate well with faults that were logged by the client. Newly identified gouge zones that correlate across multiple drillholes are suggestive of previously unidentified structures.
The results of the automated core logging were utilized in a subsequent update of the structural model of the Neves-Corvo mine. The CDI model and gouge model results were particularly useful in identifying faults and damage zones. These datasets were used to refine previously modeled structures and to model new structures in structurally complex areas or when structures were too subtle to be consistently logged and correlated across drillholes.
