The monocrystalline solar cell (MSC) interior is prone to miscellaneous defects that affect energy conversion efficiency and even cause fatal damage to the photovoltaic module. In this study, an automatic defect inspection method for MSC interior is presented. Electroluminescence (EL) imaging technology is utilized to visualize defects inside MSC. Also, accurate cell positioning is the precondition of full inspection, so a sigmoid-Hough-transform-based geometric segmentation (SHTGS) algorithm is designed to extract the complete cell region in the EL image, even though the fuzzy boundaries of the cell contain defects. Furthermore, a self-comparison method (SCM) is proposed to detect defects in the background with nonuniform luminance and complicated texture. The experimental results verify the effectiveness of this method in terms of inspection speed and recognition rate.