Underground coal mine image processing method based on deep neural network

Abstract

An underground coal mine image processing method based on a deep neural network comprises the steps of constructing an eight-layer convolutional neural network by use of an AlexNet network structure model, and completing networking training by use of a Tensorflow deep learning training framework; then, realizing training and test for the convolutional neural network subjected to initial training by use of images with five types of image quality including high brightness, low brightness, high noisy point, low contrast ratio and low resolution ratio to obtain a deep convolutional neural networkwhich can realize classification of image quality; and in combination with a mature image processing method at present, respectively adopting different types of image processing methods for images with different types of image quality. According to the underground coal mine image processing method based on the deep neural network, the content is the key problem which needs to be solved by detection visualization of a coal mine disaster area and also provides knowledge reserve and technological base for future large-scale safe mining of deep coal resources in China.

Description

technical field [0001] The invention relates to an image processing method, in particular to a deep neural network-based coal mine underground image processing method. Background technique [0002] Intelligent video monitoring is an indispensable part of the coal mine safety production system. Most large and medium-sized coal mining units and major mining research institutions have been equipped with intelligent video monitoring systems. The monitoring personnel can monitor the coal mine underground through various video facilities in a timely manner. Master the personnel situation, equipment working situation, safety situation and other information. Therefore, intelligent video surveillance provides a strong guarantee and support for the safe and smooth progress of coal mine production. However, accidents in coal mines still occur from time to time, and many accidents are identified as accidents caused by poor image quality of the video surveillance system. [0003] For t...

AI Technical Summary

Problems solved by technology

However, accidents in coal mines still occur from time to time, and many accidents have been identified as accidents caused by poor image quality of the video surveillance system

[0003] For the existing image processing methods, for different images to be enhanced, because of the difference in enhancement methods and working mechanisms, there is no efficient and universal image enhancement method, and there are a lot of problems that need to be solved urgently in existing theories and technologies

[0004] In particular, the images of underground coal mines have the characteristics of low contrast, uneven light distribution, too strong light in some areas, too weak light in some areas, high image noise caused by large dust in the production area, and obvious light spots. The imaging quality of the image is relatively poor, resulting in poor visual effect of the image

Images