Unbalanced ship classification method based on deep convolutional neural network

Abstract

The invention discloses an unbalanced ship classification method based on a deep convolutional neural network. The method comprises the steps of 1) obtaining basic data; 2) carrying out data processing; 3) constructing a model; 4) defining a loss function; 5) training a model; 6) applying the model. According to the invention, fine classification of ships and solving of the problem of unbalanced category data distribution are combined; on the basis of extracting the global features of the ship, the local features with category discrimination of the ship are introduced, so that the model learnsdetails of the ship, and the classification result is more accurate; a memory network and a fusion learning network are used for storing and emphatically learning samples which are difficult to learn; the learning of the samples is improved while the new samples are learned, the effect of rebalancing the category data distribution is also achieved, and the adverse effect on model training causedby unbalanced category data distribution in an actual scene is effectively avoided.

Description

technical field [0001] The invention relates to the technical field of computer vision image processing, in particular to an unbalanced ship classification method based on a deep convolutional neural network. Background technique [0002] The classification of ship images has always been a key research topic in the field of computer image processing. Building a robust ship image classification model is not only of academic research value, but also of great industrial application value. Today's many work scenarios, such as maritime ship monitoring, ship route tracking, and chasing smuggling ships, require accurate classification of ship images captured by equipment. [0003] At present, customs or shipping public security mainly rely on technicians to identify the types of ships. However, this manual method not only requires a large number of technicians, but also leads to low monitoring efficiency when the technicians are not in good condition. In recent years, thanks to t...

AI Technical Summary

Problems solved by technology

When the distribution of different types of data is uneven, or even some types of ship image data are insufficient, the existing deep learning methods are often not ideal.

Images