Advances in Borehole Micro-Image Processing

Abstract

Borehole microresistivity images are rich sources of geological information which (together with controls from elsewhere) contribute significantly to the description of sub-surface formations and their geometrical relationships. Processing and geological feature picking involve repetitive tasks performed manually, with some potentially subjective aspects which we seek to address via automation. Against this background it is desirable for the images to have full circumferential coverage, and for situations where this is not the case (for example in most images from wireline tools run in wells with diameters in excess of about 6 inches) a method has been developed to reconstruct the missing data using information from the morphological components that make up the measured parts. The full coverage images are the starting point for automated feature recognition whose purpose is to reduce subjectivity and increase the time available for integration and interpretation. A further advance addresses environmental perturbations and residual depth mismatches that remain after traditional speed correction, and restores depth-alignment in complex scenarios such as multiple crossing features in close proximity where prior solutions were challenged. We also address the fact that much of the information in microresistivity logs is not rendered (or rendered sub-optimally) using standard processes. This is particularly true of data from the new generation of wide dynamic range tools. Although dynamic normalization seeks to address this by changing the color scale mapping on a depth-by-depth basis, the standard implementation introduces artefacts at or close to high contrast boundaries. A new technique called dynamorphic processing includes information from a wider range of resistivity values using the same palette by constructing the image from separate images of the high and low spatial frequency components. This is the input to an additional process which uses an illumination-reflectance model to create images based on a virtual light source concept in which the position of the source is controlled to preferentially highlight geological features at particular orientations of interest. To leverage these developments we have characterized the microresistivity measurements in more detail than previously published and discovered that each button electrode has a unique response subtly different to that of its neighbors. This has led to a better understanding of borehole effects, and provides a starting point for image-driven petrophysics without the step of normalizing to an independent measurement. Each development is significant in its own right. Collectively they create images with more detail, and as such they enable additional quantitative analyses and provide a robust starting point for manual and automated analysis algorithms.

AAPG Datapages/Search and Discovery Article #90189 © 2014 AAPG Annual Convention and Exhibition, Houston, Texas, USA, April 6–9, 2014