Target detection method under low illumination based on pyramid enhanced network

The invention relates to the technical field of visual perception, and discloses a pyramid enhancement network-based low-illumination target detection method, which constructs a pyramid enhancement network, enhances an image and captures potential information in the image. The pyramid enhancement network firstly decomposes an image into a plurality of components with different resolutions through a Laplacian pyramid, and in the component of each scale, a detail processing module and a low-frequency enhancement filter are constructed to enhance the component. The detail processing module is composed of a context branch and an edge branch, the context branch carries out global enhancement on components by capturing long-range dependence, and the edge branch carries out texture enhancement on the components. The low-frequency enhancement filter obtains low-frequency semantic information and blocks high-frequency noise through a dynamic low-pass filter so as to enrich feature information. The pyramid enhancement