A computer-aided geological interpretation system for subsurface imaging

Abstract

We describe a computer-aided geological interpretation system developed for radio attenuation tomographic imaging of subsurface geology with immediate application in mining geophysics. Measurements in remote sensing are often expensive so that, in practice, the data-interpretation process is plagued by insufficient data. Inverse procedures are generally ill-posed leading to the use of regularization techniques which are mathematically tractable but rarely physically relevant to the problem. Our response to this fundamental difficulty has been to invert the normal imaging philosophy. Instead of using the data to reconstruct an image for interpretation by some domain expert, we provide a computational system in which set theoretic techniques are used to modify hypotheses provided by the expert in image form to mal<.-them consistent with the data and other constraints. This is an implementation of the hypothetico-deductive model of scientific method. While the actual image-processing algorithms such as tomographic inversion are based on sophisticated techniques, we have made a considerable effort to remove this level of abstraction for the benefit of geologists and geophysicists who are the target users. By employing X-Windows-based, graphical user interface methods, in conjunction with graphics and image-based visualization, the system can be used intuitively to expiow the possibilities offered by the data without an extensive know!< ige of the processing techniques. The application of the system to field examples from the Australian mining industry will be shown in the interactive presentation.