Digital photoelasticity is an important optical metrology follow-up for stress and strain analysis using full-field digital photographic images. Advances in digital image processing, data acquisition, procedures for pattern recognition and storage capacity enable the use of the computer-aided technique in automation and facilitate improvement of the digital photoelastic technique. The objective of this research is to find new equations for a novel phase-shifting method in digital photoelasticity. Some innovations are proposed. In terms of phaseshifting, only the analyzer is rotated, and the other equations are deduced by applying a new numerical technique instead of the usual algebraic techniques. This approach can be used to calculate a larger sequence of images. Each image represents a pattern and a measurement of the stresses present in the object. A decrease in the mean errors was obtained by increasing the number of observations. A reduction in the difference between the theoretical and experimental values of stresses was obtained by increasing the number of images in the equations for calculating phase. Every photographic image has errors and random noise, but the uncertainties due to these effects can be reduced with a larger number of observations. The proposed method with many images and high accuracy is a good alternative to the photoelastic techniques.