Service: Satellite scene stitching creates cohesive, drapable mosaics
PATCHWORK mosaics of ASTER scenes and Landsat scenes often show boundary discontinuities that can be jarring.
The patchwork effect is caused by a side-by-side tiling of scenes, a naive method that often produces a riot of stepped-looking strips and slivers (Figure 1).
FIGURE 1: Naively-tiled patchwork mosaic for the ASTER kaolinite index over southern Arizona. (The input ASTER data appearing in this write-up are level 1B ASTER data available from earthexplorer.usgs.gov.) To view a larger version, click on either the image or this text link.
These discontinuities are not only unsightly, they also represent a severe obstacle to meaningful data analysis. Trying to find quality mineral-exploration targets based on a patchwork mosaic would be a little like trying to locate the 10 tallest peaks in the world when all of the elevation statistics available involve different, undeclared reference levels.
Any limitations on extracting the underlying story from purchased remotely-sensed data means an explorer's time, money, and manpower will have been wasted — a situation that simultaneously inflates the explorer's unit costs and prompts a slide in productivity.
It's a situation that is avoidable, because our scene-stitching method converts a patchwork mosaic into a cohesive mosaic (Figure 2). Our method gives explorers access to remotely-sensed index images that are seamless.
FIGURE 2: Cohesively stitched mosaic for the ASTER kaolinite index over southern Arizona. Orange and red colors show locations possessing high kaolinite index values. To view a larger version, click on either the image or this text link.
The removal of a mosaic's discontinuities means that in the region of exploration interest, apples-to-apples comparisons are now achievable. The mosaic itself also becomes professional-looking, and so turns into an attractive graphic for use in official corporate presentations at conferences and investor roadshows.
Once the patchwork effect is removed from the mosaic, the explorer is freed up to carry out a host of rigorous computational analyses on the cohesive data, such as:
- alteration highlighting
- overlap-anomaly identification
- radial-feature detection
- linear-feature detection
The results of these kinds of analyses may reveal geological features of interest (such as alteration zones, or faults, or plutons, or lithological contacts) that can then be physically inspected via streamlined groundtruthing.
Results can be draped over topographic representations of the area — a helpful visual aid during rankings of anomalies, during targeting discussions, and during investment decision-making meetings.
Results can also be queried in conjuction with the area's topography data. For instance, it's possible to seek out locations where topographic ridges occur together with, say, anomalously high values in a certain mineral index.
The following material presents a relatively small-scale Arizona example of these kinds of analyses, which can also be run on much larger expanses of exploration ground.
Example: Ray porphyry copper deposit area in Arizona
The rest of the examples are going to show the area around the Ray porphyry copper deposit in Arizona. This area is shown as the black rectangle in the southern Arizona kaolinite image (Figure 3).
FIGURE 3: Extent of example area (black rectangle), which includes the Ray porphyry copper deposit, superimposed on the cohesively stitched mosaic for the ASTER kaolinite index over southern Arizona. To view a larger version, click on either the image or this text link.
Figure 4 shows the Landsat true-color image for the Ray area. The Ray open pit is located inside the black oval in the north-central part of the area.
FIGURE 4: Example area with the Ray porphyry copper deposit's open pit defined by a black oval, and the Tea Cup porphyry area defined by a black rectangle. This is a Landsat true-color image. (The input Landsat data appearing in this write-up are available from earthexplorer.usgs.gov.) To view a larger version, click on either the image or this text link.
The black rectangle indicates the Tea Cup porphyry area. This area was mapped for alteration and structure by Nickerson and others (2010).
The area is known to have undergone significant extension along roughly north-south–trending normal faults.
Alteration highlighting
Band ratios of ASTER and Landsat data can be used to highlight alteration around mineral deposits. The indices Fathom Geophysics most commonly uses are:
- Landsat's goethite, magnetite, clay, ferric iron, and total iron indices
- ASTER's kaolinite, chlorite (Figure 5), muscovite (Figure 6), quartz, mafic, and carbonate indices
FIGURE 5: Cohesively stitched mosaic for the ASTER chlorite index over the example area. Orange and red colors show locations possessing high chlorite index values. To view a larger version, click on either the image or this text link.
FIGURE 6: Cohesively stitched mosaic for the ASTER muscovite index over the example area. Orange and red colors show locations possessing high muscovite index values. To view a larger version, click on either the image or this text link.
Rules of thumb applying to the use of remotely-sensed data as part of porphyry-style targeting work include:
- Porphyry copper alteration in weathered terranes is typically characterized by highs in the clay, kaolinite, muscovite, goethite and ferric iron indices.
- Porphyries with unweathered rocks exposed (such as those at the Ray open pit) are more likely to be represented by highs in muscovite and total iron.
- Chlorite highs typically surround porphyry alteration.
Groundtruthing: Comparison with mapping
Nickerson and others (2010) undertook extensive fieldwork and mapped the alteration zones in the Tea Cup porphyry area.
Figure 7 shows these authors' chlorite and epidote alteration polygons displayed over the ASTER chlorite index.
FIGURE 7: Cohesively stitched mosaic for the ASTER chlorite index over the example area. Superimposed on this are polygons (black lines) representing areas where chlorite and epidote alteration were found in the Tea Cup porphyry area (black rectangle) as part of ground-based mapping by Nickerson and others (2010). To view a larger version, click on either the image or this text link.
Overall there is good correlation between the features appearing in the index map and the ground-mapped alteration zones.
Figure 8 shows ground-mapped sericite alteration polygons over the ASTER muscovite index. Again, there is good agreement. Note that some of the areas that are highlighted in the index map belong to a muscovite-bearing granitoid.
FIGURE 8: Cohesively stitched mosaic for the ASTER muscovite index over the example area. Superimposed on this are polygons (black lines) representing areas where sericite alteration was found in the Tea Cup porphyry area (black rectangle) as part of ground-based mapping by Nickerson and others (2010). To view a larger version, click on either the image or this text link.
Overlap-anomaly identification
In this example, the Landsat ferric iron index (Figure 9) has been combined with the ASTER kaolinite index (Figure 2) and ASTER muscovite index (Figure 6), as a way of emphasizing porphyry-style alteration.
FIGURE 9: Cohesively stitched mosaic for the Landsat ferric iron index over the example area. Orange and red colors show locations possessing high ferric iron index values. To view a larger version, click on either the image or this text link.
The results are shown in Figure 10, which is a CMY image that emphasizes areas where strong ferric iron, kaolinite and muscovite index values are overlapping. Coinciding highs in all three indices are rendered as black areas in this image.
FIGURE 10: Overlap analysis results: CMY imaging that combines the values of three cohesively stitched mosaics over the example area: the Landsat ferric iron index, the ASTER muscovite index, and the ASTER kaolinite index. Coinciding highs in all three indices are rendered as black areas. To view a larger version, click on either the image or this text link.
The precise level of darkness of the anomalies can be quantified, to define the extent of areas that indicate an anomalously high degree of overlap. The boundaries of these anomalies can then be vectorized to create overlap-anomaly polygons.
Radial-feature detection
Disseminated alteration found around porphyry copper deposits can be highlighted using a radial symmetry filter. The filter can be modified to highlight anomalies of different scales and can be used in an amplitude-dependent mode (to find the strongest radially symmetric signatures in the area) or amplitude-independent mode (to find radially symmetric signatures in the area regardless of their strength).
The radial-symmetry detection example shown here has been run on the muscovite index (Figure 11).
FIGURE 11: Cohesively stitched mosaic for the ASTER muscovite index over the example area (displayed using a linear color stretch to show areas with the greatest index amplitude). To view a larger version, click on either the image or this text link.
Radial symmetry results (Figure 12) were generated using a radius of 1.5 km. The analysis was run in amplitude-independent mode so that the large anomaly over the freshly-exposed rocks at the Ray deposit did not swamp out the other anomalies.
FIGURE 12: Results of radial symmetry detection run on the cohesively stitched mosaic for the ASTER muscovite index over the example area. Orange and red colors show locations possessing high radial symmetry. To view a larger version, click on either the image or this text link.
Overlap analysis, as shown previously for the index-value grids, can also be completed using radial symmetry results (Figure 13).
FIGURE 13: Overlap analysis results: CMY imaging that combines the results of radial symmetry detection run on each of three cohesively stitched mosaics over the example area: the Landsat ferric iron index, the ASTER muscovite index, and the ASTER kaolinite index. Coinciding highly radially symmetric locations across all three index mosaics are rendered as black areas. To view a larger version, click on either the image or this text link.
Linear-feature detection
In this example, normal faults in the Tea Cup porphyry area, which are known to have fragmented the alteration zones there, are auto-mapped using our phase-based edge detection method, specifically looking for structures with an azimuth of -45 to 45 degrees (Figure 14).
FIGURE 14: Structure detection results: Vectorized linear structures (black lines) with an azimuth of -45 to 45 degrees that were found via edge detection run on the cohesively stitched mosaic for the ASTER chlorite index over the example area. This same mosaic is the base map shown here beneath the structures. To view a larger version, click on either the image or this text link.
Results agree well with faults extensively ground-mapped by Nickerson and others (2010) (Figure 15).
FIGURE 15: Structure detection results: Linear structures with an azimuth of -45 to 45 degrees that were found via edge detection run on the cohesively stitched mosaic for the ASTER chlorite index over the example area. Orange and red colors show locations possessing strong linear structure. Superimposed on this are mapped faults (black lines) found in the Tea Cup porphyry area (black rectangle) as part of ground-based mapping by Nickerson and others (2010). To view a larger version, click on either the image or this text link.
Linear structures are useful to seek in porphyry-related targeting programs in another way as well — because alteration haloes surrounding ore deposits can sometimes be preferentially localized along structures such as shears, faults, or lithological contacts, to form linear alteration features. These can be detected using phase-based ridge detection.
Combining results with topography
As mentioned earlier, it's often very useful to drape the mineral indices over topography grids (Figure 16), to help visualize any recognizable associations between the current landscape surface and mineral indices (which stand as proxies for the area's geology). Auto-mapped linear features can be draped over topography too.
FIGURE 16: Cohesively stitched mosaic for the ASTER muscovite index (as shown in 2D in Figure 6) draped over the topography in the Riverside locale, which is in the eastern portion of the Tea Cup porphyry area. The locale is being viewed from the southeast, at an inclination of 35 degrees. The topography has been vertically exaggerated by a factor of 2. To view a larger version, click on either the image or this text link.
Similarly, it's also possible to use draping to examine areas where topographic ridges correlate well with mineral indexes (Figure 17).
FIGURE 17: Overlap analysis results (as shown in 2D in Figure 10) draped over the topography in the Riverside locale, which is in the eastern portion of the Tea Cup porphyry area. Coinciding highs in all three indices (the Landsat ferric iron index, the ASTER muscovite index, and the ASTER kaolinite index) are rendered as black areas. The locale is being viewed from the southeast, at an inclination of 35 degrees. The topography has been vertically exaggerated by a factor of 2. To view a larger version, click on either the image or this text link.
Recall that here in our example, data for a quite detailed area (the Ray porphyry-copper deposit region of Arizona) is being examined.
The various kinds of data analyses discussed above can also be applied to large regional-scale datasets.
Acknowledgements
Fathom Geophysics gratefully acknowledges Phil Nickerson of Bronco Creek Exploration Inc., who supplied alteration mapping and polygons for the example area.
References
- P.A. Nickerson, M.D. Barton, and E. Seedorff (2010) "Characterization and reconstruction of multiple copper-bearing hydrothermal systems in the Tea Cup porphyry system, Pinal County, Arizona", Society of Economic Geologists, Special Publication 15, 299-316.
- Fathom Geophysics Capabilities Booklet (2012) Columbus, Ohio, 48 pages.
- Fathom Geophysics Capabilities Info Sheet on Intrusion Detection (2012) Columbus, Ohio, 17 pages.