Design And Implementation Of Embedded Vision Sensor Software

With the introduction of the concept of intelligent manufacturing such as “Industry 4.0” and “Made in China 2025”,China,the United States,Germany,Japan and other countries have begun the transformation process from traditional manufacturing to intelligent manufacturing.Machine vision becomes an important technology for industrial intelligence transformation because of its automatic processing and fast speed.Visual detection is one of the most important machine vision applications.It can detect defects such as cracks,marks and stains by analyzing the image of the workpiece to achieve efficient detection.Visual detection has a wide range of applications in the automotive,pharmaceutical,food,packaging and other industry fields.This paper designs and implements an embedded vision sensor based on the actual needs of a company.The sensor integrates image acquisition and image processing.The sensor complete detection by comparing the contour,color area,width and diameter of the standard part with the above charateristics of the workpiece to be detected.The users can use the client to configure related parameters and view the processing results.The embedded vision sensor software includes embedded software,client software and intelligent processing algorithms.This paper mainly describes the management center module,image acquisition module and intelligent processing module of embedded software.The management center module is mainly responsible for the overall control of the sensor system and the interaction with the client.In order to enable the client to search and connect to the client,the management center module implements the functions,including self-discovery,heartbeat packet,tasks switch and historical results viewing.The image acquisition module is mainly responsible for image data acquisition,JPEG format encoding,and manual automatic configuration of sensor exposure parameters and sensor white balance parameters.The intelligent processing module is responsible for detecting the image to be processed by the contour tool,the color area tool,the width tool and the diameter tool,and the functions of result statistics and result output.Considering the slight deviation and rotation of the workpiece detected on the conveyor belt,the image correction function is added during the detection process.The sensor system mainly has four states,including camera configuration state,tool configuration state,tools test state and running state.The camera configuration state is responsible for the setting of the camera shooting related parameters,the tool configuration state is responsible for the setting of the relevant parameters of the tool,the tool test state is responsible for testing the currently configured inspection tool to help the user to better adjust the tool parameters,and the running state is responsible for capturing the image of the workpiece to be inspected,processing it through the inspection tool,and outputting the result through the IO and displaying it through the client software.The sensor system also designed a shared buffer area and a contour algorithm template area in memory to solve the problem of big data interaction between above modules.The function and performance of the embedded vision sensor are tested and analyzed.And the functions of the system are verified one by one according to the test case.The test results show that the function modules of the sensor function normally,and the expected performance requirements of the sensor users are also satisfied.The sensor is put on trial in actual industrial sites.