Monocular depth estimation method and system based on adaptive token aggregation

The invention belongs to the technical field of image processing, and provides a monocular depth estimation method and system based on adaptive token aggregation in order to solve the problem that accurate estimation cannot be realized due to the fact that rich global information cannot be accurately extracted and local features cannot be accurately estimated in an existing method. According to the method, respective advantages of a convolutional network and a Transform are fused and applied to a depth estimation task, the Transform is used for extracting global context information, and the convolutional network is used for retaining local context information, so that an algorithm has the capability of extracting complete information in a scene, information of Transform features and information of convolutional network features are interacted, a corresponding relation is enhanced, and the algorithm is more efficient and more efficient. And the characterization capability of the features is enhanced, so that t