A ship part decoupling identification method and device based on high-resolution range profile

Abstract

The application relates to a ship component decoupling identification method and device based on high-resolution range profile. The method comprises the following steps: obtaining a training sample set comprising a ship component echo signal, a ship overall echo signal and a ship non-overall echo signal; training an initial neural network by using the training sample set to obtain a trained ship identification neural network; inputting an echo signal of a target ship to be identified into the ship identification neural network to obtain component information of the target ship output by the ship identification neural network, and determining a ship type of the target ship according to the component information of the target ship. The method provided by the application reconstructs a target ship from a component level, constructs a double-source feature prototype of a difference domain and a single domain, can accurately decouple and identify the existence of key components from a one-dimensional mixed HRRP echo signal, and can infer the category of a brand-new ship type by identifying key components carried by the brand-new ship type even if the brand-new ship type has never been seen before.

Images