Enhanced quality borehole image generation and method

A method to modify a borehole image, the method comprising:

• obtaining a data set of borehole data from a geological field, the data set comprising historical data from previously-completed borehole operations, the historical data being grouped by basin, the historical data comprising real-time images compressed at different bit rates and corresponding recorded-mode images from the previously-completed borehole operations;
• obtaining a real-time borehole image created by field personnel, the real-time borehole image being included in a compressed data stream transmitted to a surface and decompressed in real-time to reconstruct the real-time borehole image;
• obtaining well-specific information from the borehole, the well-specific information comprising at least one of:
• geomechanics data, surveying information, real-time borehole data, or other wellbore information;
• processing the data set of borehole data from the geological field, the real-time borehole image created by field personnel, and the well-specific information to generate an enhanced-quality borehole image, the processing being performed by a deep neural network comprising a convolutional encoder, a convolutional decoder, and a deep neural network, the deep neural network comprising deep convolutional layers configured to encode significant features and denoise the real-time borehole image during a decoding phase, the deep neural network learning to bridge a gap between decompressed real-time images and recorded-mode images by learning to map two image pairs to transform features across two domains and handle compression artefacts;
• and
• displaying the enhanced-quality borehole image.