A Metallic Fracture Estimation Method Using Digital Image Correlation

This paper proposes a metallic fracture estimation method that combines digital image correlation and convolutional neural networks, based on proven theory the strain distribution of component changes when crack occurs in structure. By using correlation, achieves noncontact nondestructive sensing, as well high interference immunity. We utilize system to produce graphs reflect occurrences propagations fractures during fatigue processes. A deep residual network (ResNet) regression model is trained by correlating with corresponding lengths, so propagation condition can be estimated data from correlation. In experiment, according American Society for Testing Materials (ASTM) standards, we fabricate set aluminum specimens perform tests acquisition Finally, obtain length mean absolute error 0.0077 mm, or 0.26% measuring range. The results show precision, practicality, proposed method.