Design And Implementation Of Automatic Illumination Inspection System For Circuit Board

Space radiation is an important factor that jeopardizes the stable operation of the spacecraft.At present,the radiation damage caused by space radiation to sensitive components has been studied through the combination of ground acceleration test and effect equivalence.Therefore,it is necessary for heavy ion accelerator to cooperate with relevant detection equipment to carry out ground simulation test to detect space components.However,due to the lack of or imperfection of the supporting equipment for the simulation test of the accelerator terminal,the development of anti-radiation detection technology is restricted.For this reason,the auxiliary equipment research based on Lanzhou heavy ion accelerator is carried out,and the circuit board automatic light detection system is proposed by combining computer technology and automation technology.Firstly,based on the inspection process of the circuit board lighting inspection system,this paper design the system architecture of the inspection system and divide each functional module to build a blueprint for the design and implementation of subsequent systems.Secondly,for the interconnection between various heterogeneous devices,based on the OPC UA information modeling rules,the paper establish a robot information model,build a heterogeneous device communication system,and study the engineering implementation process according to specific hardware devices.Then,for the coordinate acquisition problem of the chip to be detected in the radiation environment,combined with the SLIC superpixel and One Cut image segmentation,this paper propose the SLIC-OneCut chip recognition and positioning algorithm,and combine with the convolutional neural network to solve the detection chip identification and positioning problem.It proposes to use the SLIC superpixel algorithm to accelerate the OneCut image segmentation process to achieve the traditional chip positioning based on image segmentation.In the deep learning model construction process,this paper produce a circuit board image data set,and based on this data set,and test two typical neural network models based on image segmentation.Based on the evaluation results of image segmentation,this paper deploy the neural network model in the engineering application.Aiming at the visualization problem of the chip identification and positioning process,this paper develop the camera application based on the industrial camera SDK to realize the visual display of the chip coordinate detection process.Finally,aiming at the friendly operability of the inspectors,the paper design a human-computer interface interactive system based on KingView.According to the inspection process of the circuit board,each subsystem is controlled to cooperate with each other to realize the inspection of the circuit board chip.