Development Of Classification System Of Tunnel Surrounding Rock Cloud Based On ResNet Mode

In this paper,based on the ResNet model of deep learning,takes the cloud computing as an important means,and uses the machine vision technology to compile the cloud classification system of the surrounding rock of the highway tunnel through the C++ language on the MATLAB software.Through deep learning ResNet model,the system is used to analyze and identify the pictures of the collected palm surface.Through a series of operations and recognition techniques,such as deep learning simulation training,palm face image classification,rock and soil separation technology,palm face fissure parameter identification,and so on,the classification index of surrounding rock classification is obtained,and then the application of AHP is established.The surrounding rock classification method of this system is applied to engineering practice to verify the accuracy of the system.Through the above technology,we have researched and developed the cloud classification system of highway tunnel surrounding rock and got some achievements.(1)through deep learning technology,machine vision technology and cloud computing technology,the classification of the surrounding rock of the highway tunnel is raised to the artificial intelligence technology.The classification results can be obtained only by the input of the picture of the palm surface and a small amount of parameters,which reduces the influence of human factors on the classification of the surrounding rock.(2)the system uses machine vision technology to identify the parameters of the palm face fissure,to evaluate the face state of the palm,and to establish the corresponding relationship between the fracture ratio and the state of the palm,without the need to obtain the parameters by the redundant instruments,and then to establish the first method of the classification of the surrounding rock with the degree of hardness,groundwater development and ground stress parameters using the layer analysis method.The classification of the surrounding rock in the system is gradually realized.(3)the method and process of obtaining the picture of the palm face are specified,and the calculation method of the length and length of the crack in the picture of the palm is given,which can accurately determine the parameters of the crack.(4)the author has successfully developed the tunnel surrounding rock cloud classification system by using the C++ language on the MATLAB software and applied it to the Yongji high speed and Chang Zhang high-speed long line tunnel.The results and the traditional methods are compared and verified.It is proved that the system can be used in the engineering practice.