Abstract

Fractures have a significant impact on production in hydrocarbon reservoirs; hence, their identification is a critical issue. Image logs are one of the best tools for revealing and studying fractures in a reservoir. However, manually interpreting fractures from these high-resolution azimuthal images is often labor-intensive, subjective, and time-consuming. This work presents an interactive, semi-automated Pythonbased workflow for efficient fault/fracture picking and structural analysis directly from Formation Micro imager (FMI) log. The workflow utilizes discrete binarization to enhance the visibility of fractures, enabling interpreters to dynamically adjust threshold and depth intervals while visualizing the image log. Fractures manifesting as sinusoids can be rapidly picked through mouse clicks, with the system automatically joining points to form complete fracture traces and determining their dip and azimuth. The interpreted fractures are instantly projected onto stereonets, allowing immediate structural interpretation and quality control. This workflow significantly reduces interpretation time while maintaining geological control and transparency, bridging the gap between manual picking and full automation. It demonstrates the potential for scalable, user-guided fracture analysis, laying the groundwork for advanced integration with automated fracture detection, fracture network modelling, and reservoir property estimation in future studies.