Experimental flame trajectories of hydrogen/air and n-butane/air spherically expanding flames were measured over a wide range of equivalence ratios at pressures of 0.1 MPa and 0.05 MPa. An experimental radius range selection method was developed to determine the critical initiation and end radii for extraction of laminar burning velocity. The results showed that both flame radius range selection and nonlinear flame stretch have significant effects on laminar burning velocity extraction. As the selected initiation radius increases, the flame radius range effect increases. Thus, the widest unaffected radius range should be processed using a proper method to reduce the extraction induced uncertainty. Additionally, the extraction method, which reflects a linear relationship between the flame propagation speed and the flame curvature, well fits the experimental data for a wide radius range when Lewis number, Le > 1. And the performances of all tested extraction methods are almost equal when Le < 1, especially for Le slightly smaller than unity. The nonlinear method is still recommended in extraction for very small Le. The critical initiation radius depends strongly on the Lewis number and initial pressure, and increases linearly with the absolute value of Markstein length. At large and small Le, the initiation radius used in measurements should be large enough to ensure reliable extraction results. Furthermore, the critical initiation radius is affected by transition and spark-assisted propagation regimes for Le > 1 and Le < 1, respectively. It is concluded that the proper flame radius range for extraction is narrowed primarily by the strong flame stretch effects, which increase the critical initiation radius for stable flames.