Model generation method, well-to-seismic joint velocity modeling method, device and equipment

The invention provides a model generation method, a well-to-seismic joint speed modeling method, a device and equipment, and the method comprises the steps: building a texture migration network model, and enabling a texture migration network to comprise a generator, a discriminator and an auto-encoder; the generator and the discriminator are used for forming a conditional generative adversarial network, and the auto-encoder is used for reconstructing the logging speed; acquiring sample data, and training the texture migration network model based on a conditional generative adversarial network formed by a generator and a discriminator; the sample data comprises an initial velocity and a depth profile; inputting the logging speed into an auto-encoder, and performing parameter optimization on the texture migration network model; the texture migration network model is used for outputting a speed modeling image according to the input initial speed and the depth profile, and the speed in the speed modeling image is