**2. Methodology**

This section provides a description of the proposed methodology. The main goal of the proposed approach is the automatic assessment of the development of concrete cracks by combining two computer vision methods: 2D digital image correlation and region-based convolutional neural network. Figure 1 shows the flowchart of the proposed approach. During the 1st and 2nd step, the tested element is prepared and the vision system is set. In the 3rd step, during the experiment, the digital images are taken and stored using DSLR cameras. In the 4th step, the stored images are pre-processed and the Faster R-CNN model is developed using the stored images. During the 5th step, the developed Faster R-CNN is used for detection and localization of surface cracks on the tested element. In the 6th and 7th step, DIC method is used to compute the displacements, strains and the crack width, respectively. Finally, the automatic assessment of the cracks during the experiments done with Faster R-CNN and DIC can be saved for the later analysis.
