Wellbore Images Digital Fusion: Beyond Single-Sensor Physical Constraints

Abstract

In the modern oilfield, borehole images can be considered as the minimally representative element of any well-planned geological model/interpretation. In the same borehole it is common to acquire multiple images using different physics and/or resolutions. The challenge for any petro-technical expert is to extract detailed information from several images simultaneously without losing the petrophysical information of the formation. This work shows an innovative approach to combine several borehole images into one new multi-dimensional fused and high-resolution image that allows, at a glance, a petrophysical and geological qualitative interpretation while maintaining quantitative measurement properties. The new image is created by applying color mathematics and advanced image fusion techniques: At the first stage low resolution LWD nuclear images are merged into one multichannel or multiphysics image that integrates all petrophysical measurement’s information of each single input image. A specific transfer function was developed, it normalizes the input measurements into color intensity that, combined into an RGB (red-green-blue) color space, is visualized as a full-color image. The strong and bilateral connection between measurements and colors enables processing that can be used to produce ad-hoc secondary images. In a second stage the multiphysics image resolution is increased by applying a specific type of image fusion: Pansharpening. The goal is to inject details and texture present in a high-resolution image into the low resolution multiphysics image without compromising the petrophysical measurements. The pansharpening algorithm was especially developed for the borehole images application and compared with other established sharpening methods. The resulting high-resolution multiphysics image integrates all input measurements in the form of RGB colors and the texture from the high-resolution image. The image fusion workflow has been tested using LWD GR, density, photo-electric factor images and a high-resolution resistivity image. Image fusion is an innovative method that extends beyond physical constraints of single sensors: the result is a unique image dataset that contains simultaneously geological and petrophysical information at the highest resolution. This work will also give examples of applications of the new fused image.