b'AEGC 2021Short abstractsAEGC 2021: Short abstracts3: Mapping coal seam roof and floor by using in-seamEM systems. Using modern electronics and software we have borehole radar: Results from numerical modelling been able to overcome many of the problems associated with the broad bandwidth needed to define near surface conductivity Dr Binzhong Zhou 1, Dr Jianjian Huo2, Prof Iain Mason3 andwith a time domain system. Sampling at around 500 000 MrMatthew van de Werken1 samples per second and processed to produce a measurement of secondary field every second, the Loupe system provides very 1 CSIRO high spatial resolution. Data can be viewed as the operators 2 Ningbo University walk, allowing survey redesign as necessary.3 The University of SydneyDuring 2019 and 2020 trial surveys have been conducted with Longwall automation depends on the development ofLoupe in a number of near-surface applications including automatic coal seam (horizon) tracking and lateral guidancemineral exploration on surface and underground, geological / systems to keep mining the target coal seam while steering toregolith mapping, study of groundwater around tailings storage the desired target. While inertial sensing has provided somefacilities and the mapping of structural features in open-cut success for lateral control of mining machinery, horizon control,mines. We see a wide application for Loupe in mapping seepage i.e. establishing a flight plan for the longwall shearer, is aboth from mine tailings and acid mine drainage.critical component of longwall automation that requires spatial accuracy for changes in a horizons depth of the order of 10 cm,The Loupe system has proved to be extremely versatile and positions of structures (faults and rolls) within fractions of aworking in difficult terrain and areas with high electromagnetic metre tens of metres ahead of the longwall machine. However,interference such as mine sites and urban sites. Special securing this is still largely unachieved. Much depends on ourchallenges are presented when working underground due precise prior knowledge of the coal seams location. to power reticulation, vehicle movement, infrastructure and particularly steel mesh reinforcing. We will give examples Coal is electrically resistive compared with its surroundingshowing data collected in these challenging circumstances.rocks. Therefore, we can use electromagnetic waves to image coal seam boundaries (roof and/or floor) and its relevantDuring this presentation, we will summarise the Loupe system structures. Here we propose to use borehole radar profiles ofand show results from several recent surveys.in-seam drill holes to map the seam roof or floor accurately. This will help to fill the gap between coarse scale and relatively10: Inferring geological features masked by artefacts in inaccurate exploration data (drilling, borehole logging andcore photography using neural networksseismic) available and the detailed seam knowledge (ground penetrating radar, thermal and optical data) at the longwallDr Yasin Dagasan 1, Mr Harvey Nguyen1,2 and Mr Mark Grujic1face and gate roads. Integrating these datasets should provide1 Solve Geosolutionsa more accurate horizon model that can be used for longwall2 Datarockmachine automation and guidance.In this paper, numerical modelling is used to investigate theComputer vision is considered to be the theory and supporting feasibility of and factors affecting in-seam borehole radartechnology to extract information from imagery or multi-imaging. We show that in-seam borehole radar imaging can bedimensional data, with the goal of automating tasks that would used to map the coal seam roof and floor position accuratelyotherwise be performed by humans.with an estimated error of less than 10 cm. Such accuracyThe application of computer-vision technologies to geological requires the central frequency of the borehole radar to be nodatasets has previously been shown to objectively verify less than 100 MHz and the offset of the transmitter and receiverexisting datasets (e.g. automated calculation of geotechnical no more than 1 m. parameters) as well as provide new datasets (e.g. fracture orientation data for an entire deposit).7: Case studies from LoupeNew technology inWhen visually inspecting core, a geologist will infer geological portable TEM for near-surface measurements features that are masked by artefacts like hand-drawn mark up; a vein is not intersected by an orientation line and the geologist Mr Gregory Street 1 and Mr Andrew Duncan2 knows this. Similarly, physical inspection of core means that interpretation is not affected by artefacts that can manifest in 1 ASEG AIG GSA photography, including shine from lighting in the core shed. A 2 EMIT,Loupe Geophysics computer vision model does not initially know the relevance of A portable, broadband TEM system, Loupe, has been developedthese non-geological artefacts and must be trained accordingly.for the purpose of measuring the distribution of near-surfaceWe introduce the application of neural network architectures to electrical conductivity. The system records continuously usingidentify and infer geological information that is hidden by non-a three-component coil receiver mounted on an ergonomicgeological artefacts in core photos, including; a) core markup backpack from signals generated from a small (660mm)such as orientation or cut lines and handwriting, b) joining rock diameter transmitter loop mounted on a similar backpack. pieces either side of mechanical/drillers breaks, and c) lighting The Loupe system is designed to measure primarily in the topeffects including shine in wet photos taken using non-diffuse 25 m of the ground, previously the charter of frequency-domainlight sources.63 PREVIEW AUGUST 2021'