The use and abuse of image analysis in geophysical potential field interpretation

Abstract

Images of geophysical potential field data are becoming more common as a result of the increased availability of image analysis systems. These data are processed using techniques originally developed for remotely sensed satellite imagery. In general, geophysicists are not familiar with such techniques and may apply them without due consideration. This can lead to abuses of the geophysical data and reduce the validity of the interpretation. This paper describes some critical processes which can introduce errors to the data. The production of a regular grid from scattered data is fundamental to image processing. The choice of cell size is paramount and must balance the spatial distribution of the data. The necessary scaling of data from real values into a byte format for display purposes can result in small anomalies being masked. Contrast stretching of grey level images is often applied but can alter the shape of anomalies by varying degrees and should be avoided. Filters are often used to produce shaded relief images but without due regard to their frequency response and the effect on images expanded to fill the display space. The generation of spurious numerical artefacts can be reduced by ensuring that the filter is applied at real precision to the original data grid. The resultant images can then be processed for display. The use of image analysis systems for data integration requires careful consideration of the sampling strategy and information content of each dataset. It is proposed that such procedures are more appropriately conducted on a geographic information system.