This thesis presents a methodology for measurement of strain fields in materials by applying the SIFT technique (Scale Invariant Feature Transform). The images analyzed are captured by a digital camera at different stages during the deformation process of a material when it is subjected to mechanical stress. SIFT is one of the modern computer vision techniques and an efficient filter for extraction and description of relevant feature points in images. These interest points are largely invariant to changes in scale, illumination and rotation. The methodology is based on the calculation of the gradient of the functions that represents the corresponding strain field in the material during a mechanical test under different boundary conditions. The functions are calibrated with the application of inverse analysis on the set of homologous points of two images extracted by the SIFT algorithm. The formulation of the solution to the inverse problem combines the experimental data processed by the SIFT and linear least squares method for the estimation of strain parameters. The proposed models for different specimens are evaluated experimentally with strain gauges for direct measurement of the deformations. The strain field identified by the computer vision system is compared with values obtained by strain gauges and simulations with the ANSYS finite element program. The proposed models for different types of measurements are experimentally evaluated with strain gages, including the estimation of mechanical properties. The results show that the strain field can be measured using the SIFT technique, developing a new visual tool for measurement of mechanical tests that are not based on traditional techniques of image correlation.