Local Intensity-based Correction for SMFT Longitudinal Magnetograms: Validation with SDO/HMI Observations

Abstract The Solar Magnetic Field Telescope (SMFT) at Huairou Solar Observing Station (HSOS) derives photospheric magnetic fields from single-wavelength Stokes measurements using a linear calibration based on the weak-field approximation. In sunspot umbrae, this linear approach systematically underestimates the true magnetic field strength---a well-known effect of magnetic underestimation. To mitigate this bias, we apply the Local Intensity-based Correction (LIC) as implemented in the standard HSOS data reduction procedure and validate its performance against SDO/HMI full-disk longitudinal magnetograms that are co-temporal and co-spatial with the SMFT observations for \textbf{26 active regions observed in 2019 and 2024}. Pixel-level alignment between SMFT images ($992\times830$, $0.242^{\prime\prime}$~pixel$^{-1}$) and HMI snapshots ($4096\times4096$, $0.505^{\prime\prime}$~pixel$^{-1}$) is achieved using SIFT feature matching and RANSAC-based homography estimation, yielding inlier fractions of 40\% to 88\%. The LIC correction consistently improves the Pearson correlation coefficient $r$ with respect to HMI for all these 26 regions, with a mean improvement of $\Delta r = +0.041$ and a maximum of $\Delta r = +0.197$ for AR24114 (NOAA 13667). Residual discrepancies persist in the most severely underestimated pixels, highlighting an inherent limitation of any single-wavelength linear calibration. Nevertheless, our results demonstrate that the LIC correction is a robust and effective preprocessing step, and we recommend its routine application in quantitative analyses of SMFT longitudinal magnetograms.