CN-122023137-A - Micro-resistance imaging logging image processing method, device, equipment and medium

CN122023137ACN 122023137 ACN122023137 ACN 122023137ACN-122023137-A

Abstract

The embodiment of the application provides a processing method, a device, equipment and a medium for a micro-resistance imaging logging image. The application relates to the field of image processing, which comprises the steps of obtaining a well logging image to be completed and mask images corresponding to the well logging image to be completed, inputting the well logging image to be completed and the mask images into an image completion model to obtain a preliminary well logging image, carrying out corrosion treatment on the mask images to obtain corrosion mask images with corrosion areas, filling pixel values corresponding to the corrosion areas in the preliminary well logging image to corresponding pixel missing positions in the well logging image to be completed to obtain a new well logging image to be completed, taking the new well logging image to be completed as the well logging image to be completed, taking the corrosion mask images as the mask images, and repeatedly executing the processing of inputting the well logging image to be completed and the mask images into the image completion model until the well logging image after completion is output. The method is used for achieving the technical effect of improving the accuracy of the full micro-resistance imaging logging image.

Inventors

FENG CHENG
CHEN JUNTAO
ZHANG GUOZENG
CHEN JUNKAI
ZHANG NING
FENG ZIYAN

Assignees

中国石油大学(北京)

Dates

Publication Date: 20260512
Application Date: 20251216

Claims (10)

1. A method of processing a microresistivity imaging log image, the method comprising: acquiring a well logging image to be completed and a mask image corresponding to the well logging image to be completed, wherein the well logging image to be completed comprises a plurality of pixel missing positions; inputting the well logging image to be complemented and the mask image into an image complement model to obtain a preliminary complement image; performing corrosion treatment on the mask image to obtain a corrosion mask image with a corrosion area; filling pixel values corresponding to the corrosion areas in the preliminary completion image to the pixel missing positions corresponding to the well logging image to be completed, so as to obtain a new well logging image to be completed; And taking the new well logging image to be completed as the well logging image to be completed, taking the corrosion mask image as the mask image, repeatedly executing the processing of inputting the well logging image to be completed and the mask image into an image completion model until all pixel missing positions of the well logging image to be completed are filled with pixel values, and outputting the well logging image after completion.
2. The method of claim 1, wherein the filling the pixel value corresponding to the corrosion region in the preliminary completed image to the corresponding pixel missing position in the well log image to be completed, and after obtaining a new well log image to be completed, the method further comprises: judging whether all the pixel missing positions are filled with pixel values or not; If yes, taking the new well logging image to be completed as the well logging image after completion; and if not, executing the processing of taking the new well logging image to be completed as the well logging image to be completed and taking the corrosion mask image as the mask image.
3. The method of claim 1, wherein the training process of the image complement model comprises: acquiring a plurality of training samples, wherein each training sample comprises a sample well logging image and a sample mask image corresponding to the sample well logging image; inputting the sample well logging image and the sample mask image into the image complement model to obtain a first complement image corresponding to the sample well logging image; Performing expansion processing on the sample mask image to obtain an expansion mask image with a larger missing area; inputting the image complement model according to the sample well logging image and the expansion mask image to obtain a second complement image corresponding to the sample well logging image; and training the image complement model based on the second complement image and a preset total loss function to obtain the trained image complement model.
4. The method of claim 3, wherein the inputting the image complement model from the sample log image and the inflation mask image to obtain a second complement image corresponding to the sample log image comprises: Setting the pixel value of the region corresponding to the larger missing region in the sample well logging image to 0 to obtain a missing image; and inputting the missing image and the expansion mask image into the image complement model to obtain the second complement image.
5. The method of claim 3, wherein after the obtaining the first complement image corresponding to the sample log image, the method further comprises: and calculating the reconstruction loss between the effective data area of the sample mask image and the first complement image according to a reconstruction loss function.
6. The method of claim 5, wherein training the image complement model based on the second complement image and a preset total loss function to obtain the trained image complement model comprises: Calculating a perceived loss, a style loss and a self-masking loss of the missing region for the second complement image, respectively; Calculating a total loss value according to the preset total loss function, wherein the total loss value is a weighted sum among the reconstruction loss, the perception loss, the style loss and the self-masking loss; And training the image completion model based on the total loss value until the total loss value converges, so as to obtain the trained image completion model.
7. The method of any of claims 1-6, wherein the image complement model comprises an encoder, a decoder, and a prediction head, wherein the encoder is configured to extract image features based on a multi-level downsampling structure, wherein the decoder is configured to transpose convolution based on a multi-level upsampling structure to recover image size, and wherein the prediction head is configured to generate a complete complement image.
8. A device for processing a microresistivity imaging log image, comprising: the system comprises an acquisition module, a correction module and a correction module, wherein the acquisition module is used for acquiring a well logging image to be corrected and a mask image corresponding to the well logging image to be corrected, and the well logging image to be corrected comprises a plurality of pixel missing positions; the processing module is used for inputting the well logging image to be completed and the mask image into an image completion model to obtain a preliminary completion image; the processing module is also used for carrying out corrosion treatment on the mask image to obtain a corrosion mask image with a corrosion area; the processing module is further used for filling pixel values corresponding to the corrosion area in the preliminary completion image to the pixel missing positions corresponding to the well logging image to be completed, so that a new well logging image to be completed is obtained; And the processing module is further used for taking the new well logging image to be completed as the well logging image to be completed, taking the corrosion mask image as the mask image, repeatedly executing the processing of inputting the well logging image to be completed and the mask image into the image completion model until all pixel missing positions of the well logging image to be completed are filled with pixel values, and outputting the well logging image after completion.
9. An electronic device is characterized by comprising a memory and a processor; The memory stores computer-executable instructions; The processor executing computer-executable instructions stored in the memory, causing the processor to perform the method of any one of claims 1-7.
10. A computer readable storage medium having stored therein computer executable instructions which when executed by a processor are adapted to carry out the method of any one of claims 1-7.

Description

Micro-resistance imaging logging image processing method, device, equipment and medium Technical Field The present application relates to the field of image processing, and in particular, to a method, apparatus, device, and medium for processing a microresistance imaging log image. Background In the petroleum and natural gas exploration and development process, micro-resistance imaging logging of a stratum is a key technology for acquiring a high-resolution resistivity image of a well wall stratum. The technology measures the resistivity distribution of stratum around the well hole through a plurality of groups of electrode arrays on the logging instrument, and generates a two-dimensional image along the well depth direction for identifying complex geological features such as cracks, layer structures, pore structures, heterogeneous interfaces and the like. However, due to the fixed electrode spacing and physical structural limitations of the logging instrument, the coverage of the instrument against the circumference of the borehole wall is often not up to 100%, resulting in the occurrence of longitudinally empty bands in the image that are continuously distributed in the depth direction of the borehole. The geological information corresponding to the blank strips is lost in the acquisition stage, and the integrity and continuity of the image are seriously damaged. In the prior art, the traditional interpolation method, the multipoint geostatistical simulation method and the deep learning method are mostly adopted for the completion method of the micro-resistance imaging well logging image. The traditional interpolation method is used for filling by depending on local pixel neighborhood relations or statistical models, is suitable for images with small missing areas and simple structures, performs complementation processing on the images through a geological model library preset in advance by a multipoint geostatistical simulation method, and performs complementation processing on the images based on a convolutional neural network and a generated countermeasure network by a deep learning method. However, in the prior art, the micro-resistance imaging logging image is influenced by factors such as geological environment, the image missing condition is complex, and meanwhile, the complete and incomplete imaging paired data of a plurality of same well sections cannot be obtained due to sampling errors in the actual logging process. Therefore, the prior art has the technical problem of low completion accuracy of the micro-resistance imaging log image. Disclosure of Invention The embodiment of the application provides a processing method, a device, equipment and a medium for a micro-resistance imaging logging image, which are used for achieving the technical effect of improving the accuracy of the completed micro-resistance imaging logging image. In a first aspect, an embodiment of the present application provides a method for processing a micro-resistance imaging log image, including: Acquiring a well logging image to be completed and mask images corresponding to the well logging image to be completed, wherein the well logging image to be completed comprises a plurality of pixel missing positions; Inputting the well logging image to be completed and the mask image into an image completion model to obtain a preliminary completion image; Performing corrosion treatment on the mask image to obtain a corrosion mask image with a corrosion area; Filling pixel values corresponding to the corrosion areas in the preliminary completion image to corresponding pixel missing positions in the well logging image to be completed, so as to obtain a new well logging image to be completed; And taking the new well logging image to be completed as the well logging image to be completed, taking the corrosion mask image as the mask image, repeatedly executing the process of inputting the well logging image to be completed and the mask image into the image completion model until all pixel missing positions of the well logging image to be completed are filled with pixel values, and outputting the well logging image after completion. In one possible implementation manner, the method further includes, after filling the pixel value corresponding to the corrosion area in the preliminary completed well-logging image to the corresponding pixel missing position in the well-logging image to be completed to obtain the new well-logging image to be completed: Judging whether all the pixel missing positions finish pixel value filling; if yes, taking the new well logging image to be completed as a well logging image after completion; if not, executing the processing of taking the new well logging image to be completed as the well logging image to be completed and taking the corrosion mask image as the mask image. In one possible implementation, the training process of the image complement model includes: Acquiring a plurality of training samples, wh