Complex geologic body fine characterization method based on dual-cycle generative adversarial network

The invention discloses a complex geologic body fine characterization method based on a dual-cycle generative adversarial network, which can realize fine-grained characterization of attribute information through a conditional variation auto-encoder generative adversarial network, introduces a conditional potential regression generative adversarial network, compulses and fully utilizes implicit coding information, and realizes fine-grained characterization of attribute information through the conditional variation auto-encoder generative adversarial network. And the information is embedded into the generation process. In order to avoid the problems of many-to-one mapping, generation mode collapse and the like, the two are coupled to form a dual-loop generative adversarial network total framework, and mapping parameters are randomly extracted in a low-dimensional vector space, so that an output distribution mode is effectively modeled and depicted, and the mapping efficiency is improved. And the uncertainty in