Abstract
The seismic intensity experienced indoors directly reflects the degree of damage to the internal structure of the building. The current classification of indoor strength relies on manual surveys and qualitative descriptions of macro phenomena, which are subjective, unable to capture real-time dynamic changes in the room, and lack quantitative indicators. In this paper, we present the Image Evaluation of Seismic Intensity (IESI) method, which is based on image processing technology. This method mainly evaluates the degree of responses from objects by identifying the percentage of movement of different types of objects in images taken before and after an earthquake. In order to further improve the recognition accuracy, we combined the camera vibration degree and the object displacement between images to correct the generated earthquake intensity level estimation, so as to achieve the rapid assessment of an earthquake's intensity indoors. We took, as an example, 29 sets of seismic data from different scenarios. We used the IESI method to evaluate the seismic intensity of these scenarios. Compared with the seismic intensity evaluation results obtained by the Post-disaster Sensor-based Condition Assessment of Buildings (PSAB) and the Image-based Seismic Damage Assessment System (IDEAS) methods, the accuracy of the IESI method was higher by more than 30%, and its accuracy reached 97%. The universality of the IESI method in different indoor scenarios was demonstrated. In a low-intensity evaluation experiment, the accuracy of the IESI method also reached 91%, which verifies the reliability of the IESI method in low-intensity regions.