Deep vertical shaft wall image denoising method

Abstract

The invention discloses a deep vertical shaft wall image denoising method. The method comprises the following five steps: step 1, constructing a denoising model; 2, designing a loss function; 3, training a denoising model by using the standard image to obtain model parameters; 4, training and testing the denoising model by using an actual image, and correcting model parameters to obtain an ELU-CNNdenoising model; 5, inputting the noisy well wall image into the ELU-CNN denoising model to obtain a denoising result. The denoising model has 28 deep layers and comprises five feature extraction modules FEM and a jump link, and the jump link serially fuses the output features of the first convolution layer and the output features of each FEM, so that the low-level features extracted by the firstlayer are continuously utilized, and the sufficient extraction of the image features is ensured. When the blind noise of the well wall image is removed, the texture features of the damaged part of the well wall can be well reserved.

Description

technical field [0001] The invention relates to the field of image denoising, in particular to an image denoising method for a deep shaft wall. Background technique [0002] About 2.95 trillion tons of coal resources in my country are buried below 1000m, accounting for 53% of the total coal resources. With the continuous development of social economy, a large number of coal mines in our country have been transferred to deep mining in the 21st century. The high-temperature and high-humidity environment of deep wells and the erosion of saline water will slowly corrode the concrete of the wellbore and well wall. The ring pressure caused by the expansion of magma and mud around the wellbore that the wellbore bears will cause damage such as bursts and bulges on the wellbore. The damage must be detected and repaired in time to effectively ensure the safety of the wellbore and prevent the occurrence of wellbore rupture. [0003] Images are an important means for human beings to ...

AI Technical Summary

Problems solved by technology

The denoising methods in the existing literature mostly use standard images to test the denoising effect. The noise contained in the actual image is not single, and it is difficult to evaluate the level and type of noise. The difficulty of denoising is much higher than removing the known noise added to the standard image. noise

Images