Research On CCD Subdivision Technology For Moiré Fringe Of Grating

Precise displacement measurement has an important role in industrial production. Displacement measurement based on the theory of grating moiréfringe has great significance due to its advantages such as high accuracy, low cost and so on. Grating moiréfringe signal is periodic signal, and it needs to be subdivided to obtain higher displacement resolution. At present there are many methods for moiréfringe subdivision which mainly focus on the orthogonal sine and cosine signals of grating outputs. The actual moiréfringe signal is difficult to achieve high multiples subdivision because of the existence of orthogonal error, amplitude fluctuation error and non-constant cycle error. So this paper presents a method utilizing CCD for collecting grating moiréfringe and all phase DFT algorithm for moiréfringe signal subdivision.CCD image sensor can obtain complete cycle of moiréfringe signal with an image acquisition. The displacement value can be calculated according to the phase difference of twice phase measurements. Signal processing can compensate for amplitude fluctuations of the collected moiréfringe signal. One signal can achieve the subdivision and identify the direction without the problem of non-constant cycle. CCD collecting moiréfringe is equivalent to sampling analog signal, which lays the foundation of subdivision processing of digital signal. All phase DFT algorithm has phase invariability and small spectrum leak. By contrast with the simulation of traditional DFT, all phase DFT algorithm has higher accuracy of phase difference calculation in the sampling of non-integral cycle and small errors in initial phase calculation of the moiréfringe signals with different initial phase, which helps to achieve high accuracy calculation of moiréfringe.The design uses TCD1501C linear CCD image sensor and AD9816 converter to convert the output signal of CCD. The driving timing of CCD and AD are generated by FPGA. And all phase DFT algorithm is carried out entirely in FPGA too. The CORDIC arctangent algorithm is used to convert the complex values of spectral into phase values. All phase DFT algorithm and CORDIC algorithm consume 26 clock cycles to output the result. The grating displacement sensor with pitch of 20μm is used, and can achieve 400 times subdivision of moirésignal with the resolution of 50nm. The system is of high integration, high reliability and so on, which has a certain application value.