Reliability of silicon photovoltaic (PV) wafers is strongly influenced by defects and residual stresses from the crystallization and wire-sawing processes. Information about defects and stress in each wafer is important for improving the solar cell efficiency. An approach is developed for characterization of defects and residual stresses on silicon PV wafers. Utilizing a lock-in photoelastic imaging technique, the infrared grey-field polariscope (IR-GFP), retardation images are generated for individual silicon PV wafers taken from industry-grown single crystal stock. Full-wafer scale retardation images are compared with band-to-band photoluminescence (PL) images from the same wafers. The lock-in photoelastic imaging allows for better identification of defects than standard band-to-band PL imaging. Analytical models of elasticity are used to generate retardation patterns for dislocations and the residual thermal stresses. The theoretical retardation profiles are compared with the photoelastic image for defect identification and residual stress analysis. The approach is capable of relatively rapid wafer imaging, automated defect detection and stress analysis, and thus may be suitable for integration as an in-line reliability control process.