Remote Sensing Mineral Exploration By Keiko Hamam Sylvia

Remote Sensing & Mineral Exploration By Keiko Hamam & Sylvia Michael GEOIMAGE Pty Ltd

Presentation Overview • Brief introduction to Satellite Remote Sensing • Resolutions – Spatial, Spectral, Temporal & Radiometric • Which satellite is best for me? • What if my area has no archive imagery? • What if my area of interest is constantly covered by cloud ? - a brief introduction to Radar Data • Case Study: The use of ALOS Imagery in Mineral Exploration – Pakistan, including • How to collect or supply GCP’s • Orthorectification • DEMs from Satellite Imagery • Accuracy of Data • Software • Questions

Remote Sensing Introduction • Satellites capture imagery as digital raster datasets • Electro-optical sensors capture energy in different electromagnetic wavelengths or bands from the visible, near infrared, short-wave infrared and thermal infrared • Different ground covers reflect or absorb energy in different wavelengths

Remote Sensing Introduction - Useful Bands Visible Blue: (0. 450. 52 um) Best penetration for clear water, poor penetration through haze Near Infra Red (NIR): (0. 77 -1. 30 um) Determines biomass content, delineates water bodies Visible Green: (0. 52 -0. 60 um) Vegetation vigour assessment Visible Red: (0. 63 -0. 69 um) Vegetation discrimination, high iron oxide reflectivity Short Wave Infra Red (SWIR): (1. 306. 00 um) Determines soil moisture content, discrimination of rock types, hydrothermal clay mapping

Remote Sensing Introduction - Resolutions • SPATIAL • SPECTRAL • TEMPORAL • RADIOMETRIC

Spatial Resolution • The smallest feature that is distinguishable on an image is determined by the spatial resolution, the XY dimensions of each pixel © Space Imaging 2003 © CNES 2004

Spatial Resolutions SPOT 5 10 -metre false colour left, 5 -metre panchromatic right © CNES

Spatial Resolutions SPOT 5 5 -metre panchromatic left, 2. 5 -metre panchromatic right © CNES

Spatial Resolutions SPOT 5 10 -metre false colour left, 2. 5 -metre pan-sharpened pseudo-natural colour right © CNES

Spectral Resolution • Spectral Resolution refers to the number of different electromagnetic wavelength bands recorded by the sensor Panchromatic or black and white imagery is acquired by a digital sensor that measures energy reflectance in one wide portion of the electromagnetic spectrum. For most current panchromatic sensors, this single band usually spans the visible to near-infrared part of the spectrum.

Spectral Resolution Multispectral imagery is acquired by a digital sensor that measures reflectance in a number of bands. Current optical multispectral remote sensing satellites can simultaneously measure reflectance in three to fourteen different bands.

Spectral Resolution RGB Landsat 321 RGB Landsat 432 RGB Landsat 543

Spatial and spectral resolutions of commonly used highand medium- resolution electro-optical satellites Quick. Bird 0. 61 m Pan (Visible to NIR), 2. 44 m Multi (B, G, R, NIR) IKONOS 0. 82 m Pan (Visible to NIR), 3. 28 m Multi (B, G, R, NIR) SPOT 5 2. 5 m Merged Pan (Visible), 5 m Pan (Visible), 10 m Multi (G, R, NIR, SWIR) ALOS 2. 5 m PRISM (Visible to NIR), 10 m AVNIR-2 (B, G, R, NIR) SPOT 4 10 m Pan (Visible), 20 m Multi (G, R, NIR, SWIR) SPOT 2 10 m Pan (Visible), 20 m Multi (G, R, NIR) ASTER 15 m 3 bands VNIR, 30 m 6 bands SWIR, 90 m 5 bands TIR Landsat 7 15 m Pan (Visible to NIR), 30 m Multi (B, G, R, NIR, 2 bands SWIR), 60 m TIR Landsat 5 30 m Multi (B, G, R, NIR, 2 bands SWIR), 60 m TIR

Temporal Resolution • Temporal resolution is defined by the revisit capabilities of the satellite For example, Landsat 5 and 7 revisit the same location every 16 days. Off-nadir viewing satellites, including IKONOS, Quick. Bird and SPOT can be programmed to revisit a location every few days.

Temporal Resolution 29 November 2001 21 December 2001

Temporal Resolution 30 December 1999 25 June 2001

Temporal Resolution Change in Band 7

Radiometric Resolution • Radiometric resolution is defined by the number of greyscale values recorded in each band by the sensor For example, ALOS, SPOT and Landsat have 8 -bit or single byte data. IKONOS and Quick. Bird have 11 -bit data.

Radiometric Resolution 8 Bit – 256 shades of grey 11 Bit – 2048 shades of grey 8 Bit imagery – suitable for GIS applications 11 Bit Imagery – suitable for Remote Sensing + Processing applications

Which satellite is best for me? Questions to consider: Regional exploration, prospect exploration or mine site planning? Amount of vegetation cover? Suitability of age of archived imagery? Availability of imagery?

Applications of high resolution imagery • Base maps for planning of prospect exploration and development work and mine site planning • Planning of access roads and utilities into remote locations • Targeting prospect areas for further exploration based on topographic features • Identification of previous exploration work • Seismic planning and field operations • Detailed identification of drainage for geochemical sampling • Production of high-resolution digital elevation models

Applications of medium resolution imagery • Regional overview of large areas • Mapping of major geologic units • Determination of regional structures • Mapping recent volcanic surface deposits • Spectral processing using Landsat and ASTER • Extensive archive of imagery, particularly Landsat • Small cost for large area coverage • Production of medium-resolution digital elevation models

What if my area has no archive imagery? Satellites available for programming: • SPOT 2, 4 and 5 • IKONOS • Quick. Bird • Radarsat

What if my area is constantly covered by cloud? Electro-optical sensors are passive imaging instruments that measure electromagnetic energy emitted by the sun and reflected off the Earth’s surface. Synthetic Aperture Radar (SAR) sensors actively transmit a radar signal in the microwave portion of the spectrum and measure the strength and other characteristics of the return signal reflected off the Earth’s surface. Because SAR is active and operates in longer wavelengths, it can acquire images through cloud, fog, haze and darkness.

What if my area is constantly covered by cloud? SAR sensors measure the roughness of the surface compared to the radar wavelength transmitted. The most common wavelengths used are L-band or 235 mm (JERS and PALSAR) and C-band or 56 mm (Radarsat, ERS and Envisat).

What if my area is constantly covered by cloud? PALSAR image of Darwin Landsat 7 image of Darwin

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan At Koh-i-Sultan, Lake Resources is exploring an extensive system of intensely altered volcanics on the margin of an extinct caldera in a Quaternary age compound andesitic stratovolcano.

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Aims: • 5 -metre DEM contours to plan access for a drill rig • Stereo hardcopy for interpretation at 1: 25, 000 scale ALOS data purchases included: • 10 -metre AVNIR-2 acquired 2 October 2006 • 2. 5 -metre PRISM triplet i. e. backward, nadir and forward-looking, acquired 17 August 2006

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan ALOS AVNIR-2 acquired 2 October 2006 Visible bands shown in blue, green, red No geometric correction applied

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan ALOS PRISM acquired 17 August 2006 Forward, nadir, backward No geometric correction applied

From raw to end product – Collection of Ground Control • Most types of raw satellite imagery require some form of geometric correction or rectification so that the imagery will correspond to real world map projections and coordinate systems • Geometric rectification improves the horizontal positional accuracy of the imagery by warping the data to match identifiable features (Ground Control Points) from coordinated imagery or airphotos, maps, vectors or d. GPS points • Each ground control point should be identifiable as a single pixel on the image to be rectified

From raw to end product – Collection of Ground Control A good spread of ground control points within each individual scene and in overlapping areas will provide a good rectification result.

From raw to end product – Rectification and Orthorectification • For areas where there is undulating topography, or if the imagery has been captured at a high angle to the vertical, or very high accuracy is required, orthorectification is necessary • Orthorectification is rectification that incorporates a digital elevation model (DEM) to correct for distortions due to capture angle and topographic relief • Orthorectification is also recommended for pan-sharpening imagery where the higher resolution panchromatic data is not captured in conjunction with the lower resolution multispectral

From raw to end product – Rectification and Orthorectification An accurate and detailed DEM will improve the internal locational accuracy of each pixel.

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan • Accurate ground control was only available for the immediate area of the caldera • Systematic orthorectification of the nadir PRISM using the Geocover Landsat 7 Pan and the Shuttle Radar Topography Mission (SRTM) DEM

Digital Elevation Models (DEMs) from Satellite Imagery • DEMs from satellite imagery are produced by in- or crosstrack stereo ASTER VNIR and ALOS PRISM (right) have in-track stereo and SPOT has cross-track stereo. The agile IKONOS satellite has a combination of both in- and cross-track stereo.

Digital Elevation Models (DEMs) from Satellite Imagery ASTER VNIR band 3 N on left and band 3 B on right showing coincident GCPs in red

Digital Elevation Models (DEMs) from Satellite Imagery Epipolar images from previous ASTER datasets, left and right

Digital Elevation Models (DEMs) from Satellite Imagery Resultant DEM before editing

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Epipolar images from backward-forward PRISM pair, left and right

Accuracy of Data • The accuracy of the final DEM or imagery is very dependent on the accuracy of the ground control in X, Y and Z space and needs to match the spatial resolution of the imagery For example, Geocover Landsat 7 Pan is a good control base for imagery with a spatial resolution of 15+ metres, as it has a quoted accuracy of +/-50 metres. System corrected IKONOS and Quick. Bird both have an accuracy of +/-23 metres, excluding terrain effects, and therefore the ground control base should have a better accuracy than this.

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan AVNIR-2 Visible Blue, Green, Red Orthorectified full scene 70 km by 70 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Nadir PRISM Orthorectified full scene 35 km x 35 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Pan-sharpened AVNIR-2 Visible Blue, Green, Red Coincident Scene 35 km x 35 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Pan-sharpened AVNIR-2 Visible Blue, Green, Red right half and AVNIR -2 Blue, Green, Red left half ~2 km by 2 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan • The systematically orthorectified ALOS nadir PRISM was used for control of the AVNIR-2 • The pan-sharpened AVNIR-2 was shifted to match supplied ground control over the caldera • The accuracy of the DEM can only be assessed using the automatically generated drainage

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Resultant DEM 35 km by 35 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Resultant DEM showing generated drainage vectors 35 km by 35 km

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Resultant ALOS DEM with contours on the left and the SRTM DEM on the right

Case Study: The Use of ALOS Imagery in Mineral Exploration, Pakistan Using the ALOS imagery and DEM, we were able to supply the required 2 -5 -metre pan-sharpened imagery, pseudo-stereo hardcopy for interpretation at 1: 25, 000 scale, a 10 -metre DEM and 5 -metre contours. In addition the data was found to be of better quality than expected and exceeded our client’s expectations.

Software • At Geoimage, we use, sell and support two of the major image processing packages, ER Mapper Pro and PCI Geomatics. ER Mapper Pro is an intuitive desktop package for the processing of raster imagery. The package allows rectification of satellite imagery and orthorectification of air photos. We use it for geocoding, image compression and general image processing.

Software PCI Geomatics is an advanced image processing package for remote sensing, digital photogrammetry, spatial analysis and cartographic editing. We use it for orthorectification of satellite imagery as it models the satellite parameters and DEM generation. For the case study, we also used PCI for production of a flow accumulation image from which vector drainage lines were automatically generated.

SPECTRAL PROCESSING OF ASTER DATA ASTER VNIR bands 3, 2, 1 in red, green, blue on left ASTER SWIR bands 5+6+7+8, VNIR bands 3, 1 in red, green and blue on right

SPECTRAL PROCESSING OF ASTER DATA ASTER decorrelated SWIR bands 7, 6, 5 in red, green, blue on left Highest predicted clay minerals on an albedo image on the right

Thank you www. geoimage. com. au