Evaluation of Image-Matching Analysis to Compute Downhole Drilling Distance for Directional Drilling Automation

Abstract Drilling technology has been rapidly evolving over the past few decades, and currently, the main focus has been on enabling drilling automation. Multiple components on a drilling rig have been automated to make drilling operations safer, more efficient, and less costly. The invention of new directional drilling tools, such as Rotary Steerable Systems (RSS), has allowed directional wells to be drilled in less time and have decreased torque and drag. Directional drilling tools, in general, use inclination, azimuth, and depth measurements to locate the bit. As a result, the directional driller can orient the bit towards the desired location. While azimuth and inclination can be measured downhole, depth measurements are obtained using the pipe tally installed on the surface. As a result, it is difficult to automate current directional drilling tools because they still depend on surface measurements, manual input, and human guidance. A continuous downhole measurement system that uses inclination, azimuth, and measured depth to determine the drilling bit's location would facilitate directional drilling automation. A downhole computer would compute all calculations in the bottomhole, allowing the bit to follow a preprogrammed trajectory rather than depending on surface instructions. Minimizing surface dependence allows smoother well paths, thus decreasing torque, drag, drilling time, and costs even further. Therefore, this paper introduces a technique that computes the drilling distance using multiple imaging sensors spaced at known distances. The sensors capture images of the same formation location at different but synchronized times. The method includes an image-matching algorithm that identifies each image's fingerprint and registers similar images according to those fingerprints. It then uses the timestamps of the matched images and the distance between the sensors to compute the local average rate of penetration and the distance traveled. Here, we explain the mechanism of the image-matching analysis and introduce the concept through experiments made on sceneries that imitate the wellbore wall. Additionally, we analyze the effect of controllable variables on the accuracy of the distance calculation.