Speech recognition method and system applied to power dispatching scene

The invention belongs to the field of speech recognition, and provides a speech recognition method and system applied to a power dispatching scene, and the method comprises the following steps: constructing a corpus in the power dispatching scene; obtaining to-be-identified power dispatching audio information; constructing a speech recognition model; the speech recognition model comprises a first acoustic model and a second language model; inputting to-be-recognized power dispatching audio information into the first acoustic model, converting the audio information into a power dispatching statement and outputting the power dispatching statement, and inputting the output power dispatching statement into the second language model to judge the reasonability of the power dispatching statement. A corpus is constructed to provide data support for research on the voice recognition technology in the field of power dispatching, the improved acoustic model and the improved language model are combined, the problem of po