The HyspecIQ project – Hyperspectral Satellite Informatics for More Efficient Exploration and MiningPosted November 22, 2014
The HyspecIQ project – Hyperspectral Satellite Informatics for More Efficient Exploration and Mining
The HyspecIQ project is the subject of AMIRA project P1147. The principal researchers associated with the project, Joseph D. Fargnoli, Pamela Blake, Tom Cudahy and Adele Seymon, will present public lectures, supported by AIG, describing the project and progress to date in Perth and Brisbane during December. Joseph D. Fargnoli is with HyspecIQ, Washington DC, USA. Pamela Blake is with Boeing Space and Intelligence Systems, California, USA. Tom Cudahy is with the CSIRO Mineral Resources Flagship, Western Australia, and Adele Seymon represents AMIRA International, Melbourne.
The lectures will be in the form of “tag-team” talks that will cover the HyspecIQ system, application opportunities and the AMIRA project.
HyspecIQ is a global geoscience analytics and remote sensing informatics business which has contracted with Boeing to develop a constellation of hyperspectral imaging satellites to be launched from 2018. The HyspecIQ system has two parts, namely: (i) satellite sensors with superior spatial, spectral and radiometric resolutions and high temporal frequency/coverage (initially a <3 day repeat), combined with (ii) “multi-modal interpretation” (MMI) processing capabilities that ingest these satellite data (as well as other geoscience spatial information) to generate highly specific (and accurate) information products. Expedited digital information products will be delivered to clients within 24 hours from image capture. The first two HySpecIQ satellites will measure over 220 spectral bands between 0.4 and 2.5 µm with a <5 m pixel and a signal-to-noise performance targeting NASA’s AVIRIS-NG . Future HyspecIQ satellite systems will be designed to sense at mid-wave infrared, thermal infrared, LIDAR and/or SAR wavelengths, depending on resource industry requirements.
HyspecIQ aims to collaborate with a team of international researchers and the resources sector (private and public) through an AMIRA International project to design an optimum suite of business-critical information products across the mining cycle, from discovery to mine closure. Potential issues include: measurement of mineral alteration footprints like white mica Tschermak substitution, alunite K-Na chemistry, clinozoisite-epidote mineralogy and chlorite Mg number; exploring in poorly accessible/or and data-poor regions; exploring in deep regolith, snow, ice and/or vegetation cover; accurate characterisation of ore/waste in open pit mines and stockpiles during mining; tracking environmental impacts such as dust sources along mining infrastructure; rehabilitation progress of mining affected lands; and measurable indicators for mine closure criteria.
Mr. Fargnoli serves as the senior Vice President for Products within HySpecIQ. In this role Joseph is responsible for ensuring that the design and development of information products will address the core business needs of the user community and in the development of the collection and processing technologies to effectively address customer mission requirements.
Joseph’s areas of technical expertise are in the design and development of remote sensing systems particularly with regards to the exploitation of hyperspectral and multispectral phenomenology and in the integration of hyperspectral and multispectral data with imagery and other forms of information from multiple modalities and sources. In particular, Joseph has expertise in the development of informatics solutions incorporating remotes sensing image science with modern analytic architectures and cloud based IT infrastructure.
Joseph holds a BS in Mathematics and MS degree in Electrical Engineering from The State University of New York, MS in Optics from the University of Rochester, an MS in Telecommunications and Computers from the George Washington University and is currently pursuing further advanced graduate studies in remote sensing informatics at the Rochester Institute of Technology.
Tom Cudahy has over 25 years of research experience with CSIRO developing capabilities that deliver mineral information to the resources community from drill core, field, airborne and space-borne systems that measure reflectance/emissivity. Tom has led numerous national and international collaborative research projects (including the Western Australian Centre of Excellence for 3D Mineral Mapping) and been involved with many national and international aerospace technology development teams (including ASTER, Hyperion, SEBASS, HyMap, HISUI). His vision is explorers and miners in Australia empowered with scalable, accurate, digital, 3D mineralogy. His career highlights include: (i) 1st civilian satellite hyperspectral SWIR mineral maps of the Earth (Hyperion at Mount Fitton, South Australia); (ii) 1st seamless digital maps of mineralogy from “fresh to space” (Rocklea Dome, Western Australia)); (iii) 1st continent-scale maps of SWIR and TIR mineralogy (Australian ASTER geosciences maps); (iv) plenary keynote at the 34IGC, Brisbane; (v) tens of thousands of airborne and satelite mineral mapping products of Australia generated by Dr Cudahy and his team downloaded by users from over 40 countries; and (vi) being awarded Australian Mining’s “2012 Explorer of the Year”. Tom has a PhD from Curtin University (1999) and a BSc (Hons) from Macquarie University (1984).
Brisbane geoscientist, Helen Coles, was recently announced as a runner-up in the Queensland Government’s Science for Solutions open data competition.
Helen works with Rio Tinto Exploration in Brisbane, specialising in GIS analysis of exploration data.
Her award was for a smartphone app design that provides farmers and other rural landusers with access to groundwater data compiled by government agencies by mapping water bores and displaying the available data for them. This information is of real benefit to landholders, particularly in areas where coal seam gas development is occurring and management of water resources is critical to ensure existing landuse is not adversely affected by gas extraction. Helen’s app provides a visual representation of actual and proposed Coal Seam Gas (CSG) well depth in relation to water tables. It is designed to: increase stakeholders’ understanding of the interaction of wells with the water table; assist with infrastructure design and assessment; and reduce potential conflicts between different stakeholder groups.
The Science for Solutions open data competition awards were presented on Friday 11 July 2014 at The Cube, Science and Engineering Centre, Queensland University of Technology, as part of the GovHack-Brisbane event launch. The purpose of the competition was to promote the use, reuse and re-purposing of science data freely available on the Queensland Government open data portal. This initiative also encouraged the creation of data visualisations, application development and other unique treatments of the science datasets provided by the Department of Science, Information Technology, Innovation and the Arts.
The competition represents an innovative way of promoting effective use of public geoscience data by government. There are two sections to the competition: one for schools and an open competition. Other entries to receive awards included:
- a website and app designed to provide clear and relevant scientific information about the many species of animals from inland, marine, coastal and waterway areas as well as prehistoric information throughout Queensland;
- a mapping and reporting application designed to assist agencies rapidly assess the risks and impact of natural disasters on agriculture and horticultural areas;
- a web-browser based application that allows users to intuitively visualise, overlay, combine, filter and perform statistical analysis on data. Users can easily access other online data sources such as Queensland Government Open Data to complement other data;
- applications to assess bushfire risk, and deliver re- or near-real time information to landholders and residents affected by bushfire emergencies;
- a data visualisation, using the Queensland Globe, designed to help resolve the competing ground water uses (uptake and discharge) by different stakeholders (e.g. mining companies, farmers, government) and consider the impacts on natural systems such as groundwater dependent ecosystems; and,
- a mobile app and visualisation tool enabling farmers to identify groundwater wells susceptible to salt table rising, and modify their management strategies to maintain farm productivity.
Congratulations to all those recognised for their entries and to the competition organisers for their contribution to promoting effective, transparent and innovative access to and use of public geoscience data.
Geophysics holding the geology to account – siting mine infrastructure
Chris Wijns, Minerals co-chair, ASEG-PESA 2015 Conference and Exhibition
Planning infrastructure for a mining project depends on knowing ground conditions over the site. The main geotechnical approaches used, drilling and test pits, suffer from severe areal undersampling. Point measurements are expected to represent large areas and engineers extrapolate and interpolate from the geological logging of few drillholes to characterise sites for various elements of mine infrastructure. Errors in site characterisation can result in costly extra earthworks or wholesale changes in infrastructure layout.
Ahead of infrastructure planning at the Kevitsa mine in northern Finland, the most pressing issue was where to locate the crusher. The mine lease was tight, and amid waste dumps, the tailings dam, and operational considerations of distance from the pit, the crusher had to be located on a solid bedrock high in order to avoid costly excavation through thick, glacial till or very fractured bedrock. Drillholes were sparse outside of the resource area, so the company turned to airborne EM for consistent coverage of the whole mine lease. Prior petrophysical logging of resistivity (Figure 1) showed a sharp contrast between the conductive overburden (glacial till) and the very resistive fresh rock below. There is very little weathering on this site, but where it exists, in highly fractured bedrock, the resistivity occupies an ambiguous middle ground. Inversion of the EM signal into conductivity versus depth served as the basis for picking an interface that would represent the transition from overburden to fresh rock. A system that could record very early time channels was chosen in order to be able to map the very near surface.
Figure 1: (Top) Representative petrophysical log of resistivity showing a large jump between overburden (glacial till) and fresh rock below. (Bottom) Resistivity statistics from the entire petrophysical database illustrating the large contrast between overburden and fresh rock, plus the ambiguous middle ground of fractured and weathered bedrock.
Figure 2 illustrates the map of overburden depth using a cut-off of 500 ohm·m in resistivity. This ensures that, according to the petrophysical statistics, highly weathered and fractured bedrock is included in the overburden category, since such ground would be unsuitable as a foundation for the crusher. All drillholes at the time of planning are also included in Figure 2, with logged depth of overburden in the same colour scale as the EM-derived depth. Darker colours indicate deeper weathering. Two facts are obvious: there were no drillholes around the crusher location at the time, and the EM shows a deepening of the overburden past 30 m, which is far beyond the 15 m that the engineers were ready to accept for excavation. Based on the existing drillholes with logged overburden to about 10 m depth, and a number of test pits that recorded overburden to no more than 5 m depth around the planned plant and crusher site, the engineers were ready to interpolate between these to assume a suitable site for construction. The EM results threw this planning into doubt, and it was obvious that more geotechnical holes needed to be drilled. These were placed as shown in Figure 3, and confirmed the EM results. Logging comments are included on the figure. The primary conclusion is that previous logging, and especially test pits, recorded the extent of glacial till up to the first instance of bedrock, whether or not this was followed by intense fracturing and weathering. In this area of very deep unstable ground, the later geotechnical holes recorded alternating fractured bedrock and clay/sand layers, which would never be seen in a test pit that stops when the excavator shovel hits first bedrock. Most strikingly, at the original crusher site, these periodic clay/sand layers and fractured bedrock persist to 50 m depth. This would have been a showstopper for the construction phase.
Figure 2: Overburden depth at the plant site derived from airborne EM data and drilling. Drillholes are coloured with the identical scale as the EM-derived depth, where black is more than 30 m.
On the strength of the EM supported by new geotechnical holes, the plant infrastructure was shifted about 150 m to the northwest, as shown in Figure 3. This brought the crusher onto shallow bedrock according to both EM and drilling results, which now satisfied the engineering team. The EM mapping exercise cost $75,000 for data collection, processing, and inversion for bedrock depth, and saved the company from placing $300M worth of crusher and plant infrastructure in a bad spot.
Figure 3: EM-derived overburden depth around the crusher with follow-up geotechnical holes and logging results that confirm the EM story. The overburden depth uses the same colour scale as in Figure 2. The final crusher location is 150 m northwest of the originally planned location.
One of the main points of education about geophysical mapping, and in particular this example, is that precise correspondence with drillhole data, on a hole-by-hole basis, is unreasonable. This is apparent in Figure 2, for example. The footprint of a single airborne EM reading is over 300 m2 at the surface (from a loop about 18 m across). The “footprint” of the core from a PQ geotechnical drillhole is 57 cm2 or 0.0057 m2. The EM reading represents a very large volume average and provides a qualitative way to interpolate between drillholes, as well as to verify how well a single hole may represent the rock volume around it. Another ubiquitous caveat on comparing geological logging to geophysical mapping is that visual geological logging is inconsistent, and the greater the number of geologists involved, the truer this is. Geophysical measurements, to their advantage as well as their detriment, are entirely consistent across space and time. (The detrimental aspect is the inability to make informed and adjustable decisions about, e.g., slightly resistive overburden vs. equally slightly resistive fractured bedrock.) In the present case of overburden mapping, the challenge is to determine what the electromagnetic data are logging versus the geological boundary required and logged for in drillholes.
Site characterisation from sparse geotechnical drilling requires a lot of interpretation, from deciding what is important to log, to deciding where to place holes and how to interpolate data between them. The consequences of faulty characterisation can be anything from annoying to disastrous, with the price tag attached. Engineers hate uncertainty, and geophysical mapping of the subsurface can offer a cheap way to reduce this uncertainty beyond what geological mapping (logging) can do. Where there is a distinct physical property contrast to map, the geophysical dataset can be used to hold the geological/geotechnical model to account and make sure it doesn’t miss what is between the drillholes. From February 14-19, 2015, the Australian Society of Exploration Geophysicists, in partnership with the Petroleum Exploration Society of Australia, is hosting the ASEG-PESA 2015 Conference and Exhibition in Perth, plus associated workshops before and after. I encourage geologists to think of case studies where the geophysics was held to account by geological observations, and submit working titles for their presentations at http://www.conference.aseg.org.au, followed by full abstracts from June 1, for an opportunity to show the geophysicists what they missed between the drillholes.
This article also appeared in the May 2014 issue of AIG News, AIG’s quarterly member newsletter.