China’s LAMOST telescope is the most powerful spectroscopic measurement telescope for studying large field of view and large sample astronomy. It combines the world’s leading splicing mirror technology with thin mirror technology for the first time, breaking through the inability of previous astronomical instruments to combine large clear apertures and wide The bottleneck of the field of view. In order to more accurately detect the movement between the sub-mirrors of large-aperture telescopes, it is necessary to install a precise sensor at the edge of the sub-mirror to detect the movement between the sub-mirrors, and then adjust the mirror displacement in time through the force actuator under the mirror to obtain more Good image quality. Due to the particularity of the mirror surface, there are strict requirements on the displacement measurement accuracy of the sensor. Its research focuses on the acquisition and filtering of sensor displacement signals. In order to extract useful digital signals from strong noises and improve the signal-to-noise ratio (SNR) of the digital signals output by the detection system, this paper proposes a dual-channel digital filtering algorithm combining improved four-entry wavelet and adaptive filtering. By improving the four-entry wavelet algorithm, the algorithm improves the reconstruction capability and linearity while taking into account the multi-resolution characteristics of the traditional wavelet transform algorithm, and ensures the continuity of the wavelet coefficients at the threshold; and the second channel collects the high The frequency characteristic noise signal is processed again through adaptive filtering, and finally a sensor displacement signal with a higher signal-to-noise ratio (SNR) is obtained. After the front-end algorithm development is completed, the displacement signal acquisition and processing system is realized through the ZYNQ-7000 development platform, including AD conversion, digital filtering, signal transmission and LCD screen display, etc., and the Gui interface program is written using Matlab to convert the displacement signal Real-time display and save records on the PC side. The entire experiment of the displacement sensor was carried out at the Nanjing Institute of Astronomical Optics Technology, Chinese Academy of Sciences. The results show that the digital signal-to-noise ratio (SNR) processed by the algorithm is 20.7% higher than that of the traditional wavelet algorithm. The root mean square error (RMSE) of the fitted displacement curve is reduced by 19.8% on average, and the running time was reduced by 50%. It shows that the algorithm is accurate and fast, and its entire signal processing system has important application significance for the research of displacement sensors between splicing mirrors of large astronomical telescopes.