Imaging Logging Blank Strip Filling Based on Generative Network

Imaging logging technology is a method of imaging physical parameters of wellbore walls and surrounding objects by observing relevant parameters in boreholes. This method can directly reflect the information of stratigraphic images. However, due to the presence of gaps between measuring pads, the generated image information cannot fully cover the entire wellbore. Therefore, there is an urgent need for an efficient and high-quality filling method to fill the imaging logging blank strips, achieving more accurate and efficient geological information detection, and helping to improve the exploration and development level of complex oil and gas reservoirs. In this paper, the depth generative model based on the improved U-Net is used to iterate a noisy image for many times, and the actual pixels in the non blank area are compared with the generated pixels, so that the generated image is highly consistent with the image to be filled, and then the blank strip is filled. This approach avoids the requirement of multiple samples and has its own adaptability for different blank strip images. The introduction of multiple modules in the model training process can greatly reduce the complexity of the model, increase the efficiency of feature extraction, and accelerate the iteration number of model training. Through the prediction of imaging logging images in actual work areas, it is shown that the method proposed in this paper for filling in the well blank strips has a good filling effect on the blank strips, and has a good description of the logging image bedding, with strong applicability.