Research On Bus-Based Visual Detection Slave Station Technology Of Linear Features

Under the background of "Made in China 2025",the manufacturing industry is gradually moving towards informationization and intellectualization.New industries such as industry,driverless,intelligent logistics and so on require higher compatibility and real-time performance of visual integration.At the same time,the application of line feature detection in disturbed and low contrast scenarios is also limited.This paper mainly studies the bus integration technology of visual slave station and applies it to online feature detection.Firstly,based on the advantages of CAN Open(Controller Area Network Open)bus real-time and open,an improved bus integration scheme is proposed.Combined with the application requirements of line structured light leveling and cloth stripe location,the overall scheme is determined,which takes OpenMV embedded camera as platform,MicroPython as visual system and CANOpen as communication protocol.The line feature visual detection method is studied.In online structured light center extraction,threshold is set adaptively by region of interest and gray histogram,and several Hough transform voting results with Gaussian cross section are compared.The problem of jitter caused by ambient light interference is solved by using "Grayscale + Laplacian" superposition voting method.In fabric stripe extraction,Canny operator is used for edge detection,statistical gradient distribution is used to determine the high and low threshold and line preview angle,and Hough transform is used to extract different contrast edges.The bus integration technology and program development method of embedded visual slave station are studied.Referring to DS402 servo equipment standard,the application layer communication protocol of visual slave station is established.The communication is realized by visual slave bus interface and data reading and writing function block of main station axis process.According to microPython grammar and line detection requirement,the visual detection program and encapsulation interface are developed at the bottom level.Finally,the master station control program is compiled to achieve the purpose of visual detection and motion control.The experiment platform is built to verify the delay of bus communication,the accuracy and performance of line detection.The experiment shows that the bus integrationguarantees the real-time performance of the system,and the stability and accuracy of line detection meet the expected requirements.