ANTI-INTERFERENCE TRAINING METHOD FOR TUNNEL LINING DETECTION MODEL, AND TUNNEL LINING DETECTION METHOD

Provided are an anti-interference training method for a tunnel lining detection model and a tunnel lining detection method. The training method includes: training a tunnel lining detection model a plurality of cyclic times by a training image set, where for each training cycle, the following operations are performed: obtaining the training image set; inputting the lining radar image in the training image set into the tunnel lining detection model to obtain a first feature map and predicted first position information, adding a disturbance to the first feature map to generate a disturbance feature map, and obtaining, based on the disturbance feature map, predicted second position information by using the tunnel lining detection model; obtaining a first loss function based on the first position information, obtaining a second loss function based on the second position information, and obtaining a third loss function based on the first loss function and the second loss function; and adjusting a parameter of the tunnel lining detection model based on the third loss function. The anti-interference training method can improve the detection accuracy of the tunnel lining detection model.