Research And Application For Subdividing For Moire Fringes

In this thesis, domestic and foreign development state and grating measurement principle of grinder measure and control devices are briefly introduced. The advantages and disadvantages of several classical moire fringe division methods are analyzed and compared with each other. Subdividing for moire fringes based on artificial neural network is proposed with the purpose of research on the practicability and subdivision precision of the moire fringe signal. By the way of theoretical analysis and experiments simulation, the method is feasible and practical. At the same time, it can improve the quality of the original signal, reduce the recommend of the orthogonality and equiamplitude, and improve the subdivision precision.This thesis briefly has introduced artificial neural network and the application of MATLAB, and in detail explained design theory of subdividing for moire fringes based on artificial neural network, three layers neural network was used, choosing a few sample points in one cycle of moire fringe signal, input of network is tangent of every sample point, output of network is the micro-displacement in a cascade spacing of this sample point, train this network and build rational neural network model using MATLAB software, achieve moire fringe interpolation with microcomputer technology. Using technology of neural network to subdividing with software, combined with multiplication octave hardware subdividing, not only increases the number of subdividing, but also improve the measurement accuracy. External counter is used to count the moire fringes, the actual measured value, which is calculated in the single chip, displayed from LED display circuit.The paper detailed the design thought of each part of hardware circuit, fundamental component and the function of realization. Provided The circuit diagram of the functional modules, introduced the design thought of main program and Each function subroutine, and given the program flow chart.Subdividing for moire fringes based on artificial neural network, micro controller unit is used to control the hardware circuit system, through neural network fine points the output signal of grating, predigest the hardware design, improved on signal track capacity, subdivision accuracy, improved the reliability of system. This way of subdivision is used in the internal circle grinder detect and control device with precise and integration, achieved the good results, By means of experiment, simulation and debug to parts of circuit, it is justified that the system is feasible, meeting the design requirements, so it has some practical application value.