Research For Extracting Ground Motion Parameters Based On Image Dynamic Recognition Technology

Earthquake is a natural phenomenon.Moderate magnitude earthquakes will have a huge impact on urban systems.Ground motion parameters can provide important basic data for response spectrum design,structural seismic design,ground motion prediction models,and provide great help for earthquake disaster assessment and seismic scientific investigation.With the advent of the 5G era,the earthquake process can be transmitted quickly and recorded in real time by densely distributed cameras,and be presented more intuitively in the form of dynamic images.If ground motion parameters can be extracted quickly and accurately by dynamic images,it will be a treasure trove of seismic data.Not only can it make up for the lack of density in the layout of seismic stations,but it can also provide a lot of support for the division of urban intensity circles.A method of extracting ground motion parameters based on image recognition technology is proposed in this paper.Based on the principle of image recognition,a vibration-image recognition method based on the base point in the field is proposed,the relative motion model of the vibration-image system is established and optimized,the dynamic equation is analyzed,and the accuracy of the method is verified under the action of different input waves,which proves that the method has certain application value and can provide another level of solution for future ground motion analysis work.The paper mainly completed the following tasks:(1)Based on the principle of image recognition,technical support for the extraction of ground motion information is provided.The method of image acquisition and preprocessing is described,the related principle of image template matching and the coordinate system conversion process are explained.According to factors such as image acquisition conditions and recognition accuracy requirements,three correction methods for image recognition are proposed.(2)A vibration-image recognition method based on the base point in the field is proposed.In order to realize the vibration-image recognition of the base point in the field,a vibration-image recognition system was designed.Three different camera-shaking table relative motion models are established,namely,a relatively static model,a relative translation model,and a relative swing model.(3)Dynamic equation analysis is performed based on the three motion models constructed in the vibration-structure identification method.Corresponding to different relative motion models,motion equations is constructed.Considering different heights and different input waves,the ground motion parameters is answered.By comparing the identified and measured ground motion parameters,the relative swing model is selected as the camera motion model.(4)The accuracy of the ground motion parameter extraction method is verified by using different input waves.The relationship between ground motion parameters and intensity is used as the basis for accurate determination of ground motion parameters.the accuracy of ground motion parameter recognition under different input waves is used to verify the ground motion parameter extraction method based on image recognition technology.