USE OF ARTIFICIAL INTELLIGENCE TECHNIQUES TO THE INTERPRETATION OF SUBSURFACE LOG IMAGES

The correct interpretation of subsurface borehole log ima- ges is fundamental in the characterisation of the subsurface hydrocarbon reservoirs. This interpretation is very complex and can be strongly affected by the experience and knowledge of the area of the geologist. The aim of this work is to create an expert system to support the petroleum geologist in his interpretation work, by using artificial intelligence techniques and image processing algorithms for the analysis of image borehole logs. During the first step of this methodology some notable image features using image processing algorithms are identifies. These features are represented by values that are managed by the artificial intelligence algorithms. The second step is the classification by a supervised learner. All the classes obtained, describe the physical properties of the entire well. These classes can help the geologist in correctly describing the properties of the rocks. The overall prototype integrates the algorithms previously used for the image processing and artificial intelligence and have been tested and validated with a number of real datasets.