When damage occurs to structures, such as an aircraft wing or rotorcraft blades, it is generally initiated locally and spreads to the entire structure. To prevent the destruction of the entire structure due to the damage occurring in a local area, it is important to detect and act on damage at an early stage through real-time monitoring of the entire structure. Techniques for detecting structural damage include modal based detection techniques and strain-based detection techniques. However, these techniques can change the dynamic characteristics of a structure by attaching a sensor to the structure and it may require a large amount of sensor to increase detecting resolution. To compensate for those shortcomings, research on displacement/strain measurement using a non-contact method, Digital Image Correlation(DIC), has been conducted. DIC is a technique to measure the deformation and displacement of a structure using images before and after the deformation of the structure. In addition, it is relatively straightforward to achieve the real-time monitoring of the structures by using the DIC equipment. In this paper, a method for monitoring the presence or absence of damage and the location of the damaged area is developed by using the Class Activation Map(CAM) network, a kind of explainable artificial intelligence technology. To generate training data, the finite element method is used to obtain the displacement/strain of a target structure. The developed CAM network can classify the presence or absence of damage by considering the strain fields. Moreover, the relevant result of the CAM network is the CAM image which indicates a damaged location of the structure. Finally, the present CAM network for damage detection will be validated by considering the DIC experimental data of tensile specimens