| With the continuous development of the Internet of Things technology,systems based on the Internet of Things technology are increasingly applied to the field of smart fire-fighting.Among them,the emerging NB-IoT technology is favored due to its advantages such as wide coverage,large capacity,and low cost.In the traditional smart fire-fighting system,most of the judgments are made by transmitting the parameters of the sensor,which has the problems of low accuracy and poor timeliness,while some systems transmit the video image through the network to the background for unified processing and judgment.Problems such as a large amount of transmitted data and high background processing pressure.Aiming at the problems of the existing system,this paper designs and implements a smart fire-fighting experiment platform that uses video images to recognize flames at terminal equipment and uploads the flame recognition results through the NB-IoT network.The main research contents of this paper are as follows:(1)The requirements of the entire system are analyzed,and the overall architecture of the system is designed according to the characteristics of NB-IoT technology.The system is mainly composed of three parts: terminal equipment,IoT cloud platform and application server.The terminal equipment and the IoT cloud platform communicate through the NB-IoT network,and the IoT cloud platform and the application server rely on HTTP protocol for communication.(2)For the flame recognition algorithm part,according to the commonly used motionbased and color space-based methods,a mutual comparison and insufficient analysis are carried out,and an improved YCb Cr color model based on brightness is proposed,which can be used in more environmental conditions to realize the segmentation of suspected flame area.On this basis,through morphological processing,while eliminating sporadic noise pixels in the segmented image,small holes formed due to obstructions and other reasons are filled in to obtain a more effective suspected flame area.Extract features such as circularity,color moment and texture of the segmented area,put these feature vectors into SVM for model training,and obtain a mature classification model.Finally use this model to determine whether there is fire in the newly input video image.(3)Aiming at the hardware part of the system,according to the characteristics of the system,the hardware platform at the terminal equipment is designed and implemented.The terminal equipment mainly includes the main control module,camera module and NB-IoT module.The main control module selects the Cambrian HBoard platform,which has abundant computing and storage resources,which can meet the requirements of flame recognition at the terminal equipment.IMX-307 is selected for the camera module,which can obtain video image information in real time,and the WH-NB75-BA communication module is selected for the NBIoT module,which can ensure the normal communication of the terminal equipment.(4)For the software part of the system,the environment for embedded Linux software development was first built.And the software development at the terminal equipment was carried out,which can realize the functions of flame recognition,flame information reporting,and instruction reception.In this case,different upload cycles are adopted to achieve rapid response to the occurrence of flames while saving energy.For the IoT cloud platform,the Huawei Cloud IoT platform is selected,and its Profile file and codec plugins have been developed,which has realized its function as a bridge in the entire system.Finally,an application server was developed to receive the information forwarded by the IoT cloud platform,and realize the control of the entire system,which can control the visual display of flame records and upload cycle for each terminal device.(5)The flame recognition algorithm is tested and compared with the existing algorithms,which reflects the effectiveness and universality of the algorithm,and can effectively realize flame recognition.For the NB-IoT communication module,a communication quality test has been carried out to ensure that the normal communication of the terminal device can still be guaranteed under extreme conditions.For the IoT cloud platform,data transmission tests have also been carried out to ensure that the terminal device and the application server can be connected normally.Finally,the function of the application server was verified.The test results showed that the system can normally realize the preset data query,instruction issuance and other functions.The system runs stably and meets the design requirements. |
PDF Full Text Request