Traffic planning method and device based on deep learning and OD completion

The invention provides a traffic planning method and device based on deep learning and OD completion, and the method specifically comprises the steps: determining a first travel path between a first target city and a second target city in a preset time period based on a server, and generating bill OD information according to multiple pieces of sub-bill information; performing calculation according to population information, economic information and distance information to obtain non-bill OD completion information corresponding to the first travel path; the server performs fusion calculation on the bill OD information and the non-bill OD completion information to obtain a fusion travel coefficient of the first travel path, and performs calculation according to the fusion travel coefficient of each first travel path to obtain travel proportion information between the corresponding first target city and each second target city; and the server obtains a plurality of pieces of traffic planning information accordin