| With rapid development of intelligent manufacturing industry,machine vision technology is widely used in detecting,measuring,recognizing and positioning.These tasks can not be complete without the common support of image acquisition,image processing and embedded system,which make the algorithm development of machine vision system be more complex.Therefore,shortening the development cycle of algorithm and improve the usability,generality and flexibility is the purpose of researching the interactive algorithm development platform for embedded machine vision.In order to realize rapid development of embedded machine vision system,graph programming is used to encapsulate common algorithms and device commands into elements.Developers can complete algorithm integration by dragging,connecting and parameters configuration,and then download and operate with one key,realizing the cross-platform and real-time operation.With simple configuration,it can be used in different situation,not only suitable for junior developers with weak code base,but also for senior developers.In order to improve the usability of the algorithm development platform and make the algorithm development process more visual,the elements are constructed,which integrates parameters region,preview panel,inputs,outputs,status and running time.It not only realizes the real-time preview and single step debugging of algorithm results,but also facilitates the control of the whole program by developers.In addition,based on the analysis of the running principle of the existing graphical program,the signal triggered concurrent drive mechanism is proposed,which speeds up the responsing of graphical program and enhances the real-time algorithm debugging.In order to improve the generality and flexibility,the platform provides a user-defined interface to complete the view,modification and addition of source code.The combination of graphic and text programming can meet the demands of algorithm development in large-scale industry.In addition,combining the diversity of application fields,the platform not only encapsulates common device instructions,image processing elements,but also logical elements,machine learning elements and optimized elements.Logic elements are used for dynamic optimization of graphic programs and provide assembly structures such as condition selection and loop.Machine learning elements are used to improve the accuracy of the algorithm,meeting the requirements of high precision and intelligent system.The optimized elements are used in the enhanced processing of images,including image filtering,edge detection and other algorithms.In order to verify the feasibility of the interactive algorithm development platform,taking the thread external parameters measurement and X-ray prohibited identification as practical engineering applications.Intergrate graghic program PC-side respectively to show the assembly structure of general visual algorithm and machine learning algorithm,test code downloading and results viewing and analyze the efficiency of algorithm development. |
