Summary
This application shows how a GIS in combination with geological data sets can be used to solve specific geological problems. The training on digital image processing focuses on the usage of field and laboratory spectral data to gain a better understanding of Remote Sensing products. Advanced image processing techniques are introduced and applied using Landsat Thematic Mapper data of the Ronda-Malaga area. The Ronda-Malaga area is located in the southern part of Spain, West of the city of Malaga and Southeast of the city of Ceville.
Mineral exploration
Geologically the Ronda-Malaga area is part of the Betic Cordillera foldbelt stretching from Alicante to Cadiz in southern Spain. By combining the information that is available in digital geological maps and in attribute tables it is possible to create new maps. In this exercise two maps are created to guide in mineral exploration:
- A map showing the possible locations of copper mineralizations, and
- A map showing the potential gold deposits.
It is known that copper mineralizations occur along faults within limestones of Jurassic age and that the Permian sandstones in areas adjacent to Ronda could host gold mineralizations when they occur along normal faults. You can get further information, such as the total area or the average size of the outcrops by using Aggregation functions on the histograms of these two maps. Finally, you can analyze the density of lineaments with respect to the lithology creating a lineament density map with the Segment Density operation.
Working with the Digital Terrain Model
In geology a Digital Terrain Model (DTM) is often used to investigate the geomorphologic characteristics of the terrain in relation to the underlying influence of geology. The DTM for the Ronda-Malaga area is created by first producing a sub map of the contour segment map and subsequently interpolating the digitized contour lines of the sub map. Based on the DTM a slope map can be calculated by applying the gradient filters DFDX and DFDY and a MapCalc formula to the DTM. The slope map can for example be compared with the geology and the fracture pattern.
A pseudo relief image can be created using artificial illumination of the DTM. This technique is known as hillshading. In ILWIS hillshading is done by using the standard Shadow filter creating a pseudo relief map with artificial illumination from the Northwest or by creating a user-defined filter that simulates a different illumination.
In ILWIS it is also possible to create real three dimensional perspective views of the terrain. In this case the DTM is used as height map over which other maps (e.g. geology, lithology) can be draped.
A pseudo relief image can be created using artificial illumination of the DTM. This technique is known as hillshading. In ILWIS hillshading is done by using the standard Shadow filter creating a pseudo relief map with artificial illumination from the Northwest or by creating a user-defined filter that simulates a different illumination.
In ILWIS it is also possible to create real three dimensional perspective views of the terrain. In this case the DTM is used as height map over which other maps (e.g. geology, lithology) can be draped.
Image enhancement
In order to improve the visual appearance of the TM data various standard image processing tricks can be performed. Applying the stretch function, and some smoothing, sharpening and gradient filters on the TM bands enhances the contrast of the images.
Selecting three bands for display in a color composite image is a tedious and time consuming business since many combinations have to be tried certainly when working with ratio images.
The Optimum Index Factor technique (OIF) may help to overcome this problem. High OIF values indicate bands that contain much "information" (e.g. high standard deviation) with little "duplication" (e.g. low correlation between the bands). By using the OIF method, three band color composites can be evaluated on their effectiveness for display.
Selecting three bands for display in a color composite image is a tedious and time consuming business since many combinations have to be tried certainly when working with ratio images.
The Optimum Index Factor technique (OIF) may help to overcome this problem. High OIF values indicate bands that contain much "information" (e.g. high standard deviation) with little "duplication" (e.g. low correlation between the bands). By using the OIF method, three band color composites can be evaluated on their effectiveness for display.
Color composite TM456Spectral recognition of surface reflectors
The pixel information window can be used to investigate the DN values of the TM bands simultaneously in order to find the spectral responses of unknown ground cover types. For these unknown ground cover types you are asked to say what they are likely to be comparing their reflectance characteristics to the materials in the table containing laboratory spectral data.
Rationing
A common problem with RS images is the effect of varying illumination caused by topography. Relief causes some slopes to be illuminated more than others, thus surfaces with homogeneous reflectance properties will show varying DN values across a scene. Ratio images provide a means of correcting these differences. Creating ratio images is done using the following general formula in MapCalc:
Ratio(Tmi/TMj) = (Tmi/TMj) * 127
The importance of ratio images is that they map a spectral gradient and can therefore be used to map e.g. iron content, clay content, chlorophyll and water absorption features.
Green Vegetation Index
In order to mask vegetation it is often useful to calculate a green vegetation index. The most commonly used vegetation index is the Normalized Difference Vegetation Index (NDVI). The NDVI is defined and calculated with MapCalc as NDVI = (TM4-TM3) / (TM4+TM3) and can be used as threshold to mask vegetation in the TM bands.
NDVI
Multispectral classification
Techniques making use of training data sets are referred to as supervised classification. In order to train the classifier to perform a multi-spectral supervised classification you will have to sample the image to obtain a set of training pixels which serve as "an example of what to look for" with the classification algorithm. The output is a thematic map with the classes water, urban, limestone, kaolinite, hematite, green vegetation and dry vegetation.
Classified map of the Ronda-Malaga area
