Electroluminescence (EL) imaging is affected by off-axis illumination together with sensor and lens imperfections. The images’ spatial intensity distribution is mainly determined by the vignetting effect. For quantitative EL imaging, its correction is essential. If neglected, intensities can vary significantly (>50%) across the image. This paper introduces and tests four vignetting measurement methods. The quantitative comparison of different methods shows that vignetting should be characterized preferably in plane by the source of the same type as the photovoltaic (PV) device to be tested. A direct PV-based measurement in short distance with spatial inhomogeneity correction is proposed for general-purpose vignetting characterization. For precise vignetting characterization, vignetting-object separation using pattern recognition is proposed. The use of non-PV light sources for vignetting characterization can cause vignetting overcorrection and can even decrease the quality of the vignetting-corrected images.