Machine Vision Oriented Embedded System Development

Modern technology is developing rapidly.We are more and more intelligent to machine vision equipment.But for the current machine vision control system based on the traditional computer structure of complex and huge,the system is greatly increased in the cost and complexity of the system as a result of the volume,and the integration of the system is very low,the installation layout is very inconvenient in the production line.The embedded machine vision system,which is represented by intelligent camera,is flexible,but it can't realize complex algorithms and high precision,there are many defects in the big data processing.The design of embedded machine vision system based on the deficiency of vision system is proposed.Firstly,the Chief River Intel platform and the design index of the product are analyzed in detail.Then we design the main functional modules and the selection of the chip,especially the POE function unit.Focusing on the design of the circuit schematic diagram and the selection of components and packaging,the schematic diagram of the design of the PCB stack structure and components of the layout and wiring,the signal integrity and power integrity system simulation analysis of electromagnetic compatibility problems existing in the design of the system,such as ringing,reflection,crosstalk and simultaneous switching noise signal problems put forward specific solutions.At last,the system is simulated and analyzed by using the HyperLynx simulation software,and the crosstalk reports show that the crosstalk is less than 150 mV which is meeting the design requirements.The simulation analysis of the power supply level shows that the pressure drop is less than 5% which meets the standard of the pressure drop.The measured electric field intensity is lower than the FCC and the CISPR standard A Class and the B Class limit value of the electromagnetic radiation simulation analysis.The simulation results show that the visual system is fully consistent with the design requirements of the performance index.