Research On Large Data Compression Method Of Coal Mine Safety Monitoring Based On Compressed Sensing

With the application of Internet of Things technology in coal mine safety monitoring system,a variety of monitoring systems constructed by a variety of sensors realize the monitoring of underground personnel,equipment,environment and safety production information in coal mines,resulting in a large number of coal mine safety monitoring data,which brings great pressure to the processing and storage of servers.These massive data contain a lot of valuable information to be excavated.How to effectively compress the massive data and reduce the storage pressure of servers is an urgent engineering application problem to be solved in coal mines.Therefore,how to use effective data compression methods to compress these massive data information is the focus of this paper.This paper combines traditional compression with compressed sensing to study the compression method of large data of coal mine safety monitoring.The main research contents and achievements are as follows:(1)The data dictionary of coal mine safety monitoring is constructed.The data are pre-processed before compression.The basic data characteristics,such as time correlation and spatial correlation,are analyzed.Based on the national and enterprise standards,the data characteristics are analyzed and compared in detail,and their respective data dictionaries are established.In view of the sparse representation of coal mine safety monitoring data,by comparing with other transform bases,it can be seen that the redundant dictionary construction method based on KSVD can make full use of the characteristics of enterprise coal mine data to sparsely represent coal mine safety monitoring data,which can be better applied to the compression of a large number of data of coal mine safety monitoring,and provides an experimental basis for later research.(2)A non-linear Delta data compression algorithm based on principal component analysis is proposed.In the process of data compression of coal mine safety monitoring single sensor,the data compression algorithm of non-linear Delta error technology has the problem of inadequate threshold selection.According to the different alarm level/sensitivity,this algorithm uses dynamic soft threshold to automatically identify threshold and negative feedback tracking,effectively compresses non-abnormal process data and retains sensitive data with security hidden danger characteristics.It overcomes the hard threshold problem of Delta.The experimental results show that the algorithm has good adaptability and lays a foundation for further research.(3)A joint sparse reconstruction method for large data of coal mine safety monitoring is proposed.In the large data compression processing of coal mine safety monitoring,the basic theoretical framework of distributed compressed sensing applied to data compression and reconstruction of multi-sensor monitoring is analyzed,and the joint sparsity between multi-sensor monitoring data is analyzed.Based on the simulation platform of MATLAB,the redundant dictionary constructed in Chapter 3 is used for feature extraction,and the compressed data of the improved principal component analysis-based non-linear Delta algorithm is selected as the experimental data to compress and reconstruct a large number of data of coal mine safety monitoring.The experimental results show that the original data can be effectively recovered under the condition of large compression ratio.