Full-field Strain Measurement of Materials Using Meshless Methods and Computer Vision

The objective of this work is to provide an experimental numerical technique that combines the use of computer vision through the SIFT technique (Scale Invariant Features Transform) with a meshless numeric method to determine displacement and strain fields on the surface of a deformable material. The experimental part of this technique is based on the capture of images during the deformation process of the material when it is subjected to mechanical forces. At this stage, the SIFT technique is used as an efficient filter for extraction and description of points with relevant characteristics in the images collected. For this purpose, the material surface was previously characterized using a local application technique that is able to generate points of interest which can be detected by the SIFT algorithm as potential keypoints. By tracking successfully matched SIFT points between two images, i.e. the reference image and the deformed image, it is possible to obtain the experimentally measured displacement field. Finally, the application of the meshless formulation generates a numerical approximation for the displacement field and, therefore, the corresponding strain field can be estimated. Both solutions are built entirely in terms of these points provided by SIFT, called nodes. For this, the method of moving least squares is used. An experimental test is performed showing reliable results with good precision and robustness, proving to be a promising technique to further investigations in the field of structural mechanics.