Research And Implementation Of Real-time Monitoring System For Underground Space Environment

Underground space is relatively closed,the air can't natural circulation and the stream of etc.,makes the thermal environment of the underground space is mainly rely on equipment such as air conditioning,ventilation,to build and how to make the operation of the air conditioning,ventilation and other equipment with the best cost effectiveness,environmental parameters real-time acquisition is equipment indispensable to run efficiently.Therefore,real-time monitoring of the underground space environment study,developed a kind of based on embedded technology and computer vision processing technology of real-time monitoring and control system,the system can collect real-time underground space of the main environmental parameters and data such as traffic,air conditioning,ventilation and other equipment for underground space to implement efficient strategy provides the basis and data support.The main tasks accomplished are as follows:(1)Scheme design of real-time monitoring system for underground space environment and establishment of software and hardware platform.On the basis of investigating the requirements of underground space environment monitoring,the advantages and disadvantages of existing monitoring systems are analyzed and compared,and the overall scheme of underground space environment real-time monitoring system is established,and the software and hardware platforms are built according to the overall scheme.(2)Research and implementation of data acquisition subsystem and data processing subsystem.The collection of multiple environmental parameters is realized on the NodeMCU platform,and multiple environmental parameters are published to the Internet of things platform through MQTT protocol.On ARM platform,V4L2 video acquisition interface under Linux is used to control and set video acquisition and related parameters of USB camera by writing V4L2 application program,realizing image video data acquisition of underground space.The collected image video data were used to carry out real-time statistics of the flow of people using the related algorithms of face features and human head features.The detection of face features was realized by combining haar-like features with Adaboost algorithm.HOG feature and SVM algorithm are used for human head feature detection.Through the comparison of the statistical results of pedestrian flow between the two algorithms,the detection rate of face feature detection classifier is 81%,while that of human head feature detection classifier is 74%.Based on a large number of experiments,the algorithm based on facial features was improved,and the accuracy of the improved traffic statistics reached 95.6%.(3)Design and construction of Web server system.First,transplant BOA server,through the transplantation of CGIC library and writing CGI application to expand the server function,then design HTML pages and CGI application loaded into the page source code,realize multiple environment parameters,images,video data query and display function on the browser.The test results show that the real-time monitoring system of underground space environment can collect the main environmental parameters and human flow data of underground space in real time and meet the requirements of real-time monitoring of underground space environment.