Acquisition and Processing Report of the MULSEDRO EIT RM Project from the Otanmäki Fe–Ti–V Deposit, Finland: UAS and Ground-Based Magnetic and Hyperspectral Investigations in 2018

MULSEDRO is an EIT RawMaterial project of a consortium of Finnish, German, Danish and Swedish institutions and companies that aims to develop multi-sensor drone systems for mineral exploration. In September 2018, a field campaign was performed to test the newly designed UAS systems at
the Otanmäki Fe–Ti–V deposit in central Finland. The UAS systems are able to gather high-quality geophysical and remote sensing data relevant to the
mining industry within short time and in a noninvasive manner. The main target was a cleaned bedrock exposure (Metsämalmi area) on the premises of the former Otanmäki mine, which produced iron, titanium and vanadium products until 1985. The company Otanmäki Mine Oy plans to re-open the mine in the near future, and in this context, the collected datasets could improve the precision of distribution mapping of ore lenses as a basis for
a better ore grade estimation. The UAS surveys comprised hyper-/multispectral,
magnetic, photogrammetric and LiDAR surveys from fixed-wing and/or multi-copter drones. In addition, a ground magnetic survey was acquired,
in situ measurements with hhXRF and VNIR-SWIR spectrometers, as well as magnetic susceptibility measurements, were performed along a number of scan lines, and rock samples were taken for laboratory measurements. These and legacy ground-based data collected from the mining companies enabled the linking of magnetic anomalies and features in the hyper-/multispectral
image mosaics to the outcropping lithologies and variations in ore grade.
The fixed wing magnetic survey was conducted by Radai Oy, which covered an area of 1.14 km2 (inline/crossline spacing: 40 m/40 m; height above
ground: 40 m) within a flight time of less than one hour. In doing so, the complete eastern part of the deposit could be mapped and strong magnetic
anomalies could be associated with the magnetite-rich ore occurrences, which follow a synclinal structure.
In contrast, multi-copter magnetic surveys, acquired by HZDR-HIF, only focused on the smaller Metsämalmi outcrop area (size: ~100 x 300 m) and
were repeated with different flight altitudes and line spacings. The lowest survey was performed at a nominal flight height of 15 m and with a dense
survey line spacing of 7 m and 20 m for inlines and crosslines, respectively. The data provided comparable resolution to the ground magnetic survey.
The UAS magnetic survey took less than one hour, while the ground survey took three field days. Although this was one of the first surveys with this
setup, a proper acquisition strategy and a processing scheme were already defined. The noise level was kept acceptably low and artificial signals (e.g.
from electric motors) were effectively reduced such that the resulting magnetic data were of generally good quality. High magnetic anomalies from
the low-altitude UAS (and ground magnetic) surveys correlated well with the locations and shapes of high-grade ore lenses in the outcrop area. This
demonstrates the potential of low-altitude operating UAS to map near-surface magnetite-rich ore bodies. An equivalent layer modelling (ELM) procedure was jointly applied to all the available magnetic datasets to evaluate the consistency of data from the different magnetic surveys and to present magnetic surveys with very different acquisition parameters (different flight heights, line
spacings and inline samplings) in one consistent total magnetic intensity plot. The results of this first integration test with ELM are very promising,
but some further adjustments (i.e. finding proper weighting factors to balance individual datasets in the inversion) are required to use ELM as a standard
procedure for merging magnetic datasets in the future.
The hyper- and multispectral surveys were acquired and processed by HZDR-HIF. The data were carefully pre-processed using in-house software
routines to account for topographic and illumination distortions. Coverage with lichen and other vegetation in the outcrop area clearly reduced the
surface portion (by up to 30%) for which reliable information on surface mineralogy could be obtained. Therefore, the most valuable results were
achieved from a rather small areal of 30 x 10 m in the eastern part of the Metsämalmi outcrop, where the surface was cleaned with a high-pressure
cleaner prior to data acquisition. The multispectral survey was performed with a
fixed-wing UAS and a full-frame camera with only 4 bands in the visible and near-infrared part of the electromagnetic spectrum. The hyperspectral survey
was performed with a multi-copter UAS and a full-frame hyperspectral camera having a higher spectral resolution in the visible and near-infrared part of the electromagnetic spectrum (50 discrete raw bands from 504–900 nm). The resulting grid sizes of the multi- and hyperspectral orthomosaics were in the range of ~10 cm and ~4 cm, respectively. In addition, a photogrammetric survey
was flown with an RGB camera mounted on a fixed-wing UAS. The images collected during the survey were used to generate a digital surface
model (DSM) and orthomosaic. The resulting grid cell sizes of the orthomosaics were in the range of ~15 cm and ~7 cm for the multi- and hyperspectral
survey, respectively. For both the multi- and hyperspectral orthomosaics,
an MNF transformation was applied to enhance the spectral differences. Band ratios were calculated to visualize areas with increased surficial
iron alterations. Due to the much higher spectral resolution of the hyperspectral data, their results can be considered as more precise for estimating
iron alteration mineral abundances compared to the multispectral data. For the hyperspectral images, unsupervised and supervised classifications
were applied to distinguish pixels that were predominately associated with iron mineralization and the gabbroic host rock.
Comparison of the results from low-flying UAS magnetic and ground magnetic surveys with ground-based susceptibilities along the scan lines
revealed that the distribution of surface susceptibilities and TMI values were correlated. Strong susceptibility variations within the ore and between
the ore and host rocks led to high lateral magnetic contrasts visible in both the TMIs and the vertical derivative maps from all UAS and
ground-based surveys. Careful survey planning with a properly selected line spacing and line direction that also considered information on local
1 To obtain better correlations between the magnetic responses and the susceptibilities, inversion or modelling strategies could be an option to determine susceptibility estimates along the surface (or model the magnetic responses on the basis of the susceptibility measurements). However, such approaches are also limited due to the low resolution of the magnetic method and the large ambiguities of magnetic modelling. geology and morphology was essential to map the anomalies present with a high resolution. In the same way, areas where multi- and hyperspectral methods detected iron alteration coincided with locations where oxide ore lenses were identified from conventional surface mapping. The hyperspectral handheld spectra and the UAS HSI data at the same locations displayed a correlation. Finally, comparisons of handheld XRF and magnetic susceptibility measurements indicated that there was a distinct correlation between magnetic susceptibility and the iron oxide content. However, the relationship between induced magnetization, susceptibilities and the local magnetic field values from low-altitude UAS and ground-based surveys is non-trivial due to the strong variations in magnetic properties over short distances and the smoothed magnetic response that is associated with integration over a larger volume1. Nonetheless, the consistency of all these results indicates that it is possible to combine all the data in such a way that spatially uniform ore grade estimates along the Metsämalmi outcrop surface can be derived. Note that the main objectives of this report are to describe the multi-sensor UAS systems and present the data acquisition and processing strategies, as well as to provide preliminary results. Because some of the systems have recently been developed (e.g. a multi-copter equipped with a magnetometer is still in the trial phase), a focus is on evaluating the reliability of the data acquisition and the quality of resulting data. The overall aims of the MULSEDRO project have been described by Heincke et al. (2019). More comprehensive and concise interpretations of UAS tests in Otanmäki have been published by Jackisch et al. (2019).