Welcome to Metal Mine! Today is Share:
×

扫码分享

Modern Mining ›› 2024, Vol. 40 ›› Issue (04): 10-.

Previous Articles     Next Articles

Application of Intelligent Algorithm Based on Machine Vision in Mine Safety Monitoring System
LU Shaobo

In the mine safety monitoring system,the traditional manual detection method has been dif⁃
ficult to meet the needs of modern mines.By discussing the application of intelligent algorithm based on ma⁃
chine vision in mine safety monitoring system,the key technical requirements of mine safety monitoring sys⁃
tem are analyzed,and the development suggestions are put forward.It is found that the current intelligent al⁃
gorithm based on machine vision is mainly limited by environmental conditions in the application of mine en⁃
gineering,and it faces challenges in real-time detection and accuracy of data processing.A multi-modal sen⁃
sor is proposed to obtain more comprehensive data to enhance environmental perception.Lightweight network
models such as MobileNets and ShuffleNets are added to the model to reduce the number of model parame⁃
ters.By using the pan-connected compatible cooperative control system,multiple algorithms are run at the
same time for comprehensive analysis to enhance the generalization ability of the system.
  

  1. Shanxi Fenxi Mining Co.,Ltd.
  • Online:2024-04-15 Published:2024-08-12

Abstract:

In the mine safety monitoring system,the traditional manual detection method has been dif⁃
ficult to meet the needs of modern mines.By discussing the application of intelligent algorithm based on ma⁃
chine vision in mine safety monitoring system,the key technical requirements of mine safety monitoring sys⁃
tem are analyzed,and the development suggestions are put forward.It is found that the current intelligent al⁃
gorithm based on machine vision is mainly limited by environmental conditions in the application of mine en⁃
gineering,and it faces challenges in real-time detection and accuracy of data processing.A multi-modal sen⁃
sor is proposed to obtain more comprehensive data to enhance environmental perception.Lightweight network
models such as MobileNets and ShuffleNets are added to the model to reduce the number of model parame⁃
ters.By using the pan-connected compatible cooperative control system,multiple algorithms are run at the
same time for comprehensive analysis to enhance the generalization ability of the system.

Key words:

processing