Design And Implementation Of Lightweight Workflow Platform Based On Microservice Architecture

With the growing maturity of image recognition algorithms,the application of embedded video surveillance system is more and more widely applied,such as square traffic monitoring,airport security inspection,large outdoor scene monitoring and so on.By extracting common applications of these applications to embedded systems,it is of great research and application value to develop a general monitoring and identification embedded system.The main work of this thesis is based on the needs of intelligent monitoring of the growth status of agricultural flower fields to design a general monitoring and identification embedded system for large-scale outdoor scenes.In addition,this thesis uses this system model to design the flower growth state monitoring system based on image recognition.In this thesis,based on the characteristics of outdoor surveillance scenes and general design principles,a universal identification and monitoring system framework has been designed,which includes general hardware platform and software platform structure.Moreover,in the universal monitoring and identification system,a corresponding solution is proposed for the core problems in the model.Firstly,this thesis uses higher real-time image recognition algorithms and uses OpenCL to accelerate the recognition algorithm to solve common image recognition issues that the task's real-time performance cannot be guaranteed in embedded systems.Secondly,this thesis uses hardware coding to optimize the video compression process and uses the GStreamer framework to simplify the video monitoring development process.In addtinal,the entire design of the complex multitasking system is completed by the method of task assignment,real-time analysis,designing task-to-task communication methods and designing task real-time guarantee methods on the core tasks in the system.In the end,this thesis divides the general-purpose communication interface into different types of interfaces and designs the input interface,data flow encapsulation format and network transmission protocol for each type of interface to complete the design of the entire general-purpose communication interface.Base on these desgns and through the analysis of the specific scene characteristics and application requirements of the flower plantation garden,a robot-based intelligent monitoring system for flower growth status under this scenario has been designed,including the system framework design and the system hardware and software platform design,and has completed system design and implementation of the floral flowering period identification module,image acquisition module,transmission module,remote control mobile module,single chip microcomputer control mobile module and other core modules.