Guided filter and auto-encoder-based SAR target recognition algorithm

Abstract

The invention discloses a guided filter and auto-encoder-based SAR target recognition algorithm. Aiming at the problem that the time is consumed as the recognition carried out on synthetic aperture radar (SAR) image targets via neural networks pursues high recognition rate and complex structures are designed, the algorithm applies a rapid image fusion technology to feature extraction of an SAR recognition technology, and uses a weighted guided filter (GF) to carry out two-scale fusion on the SAR images so as to generate one-dimensional image vectors, uses an auto-encoder to carry out low-dimensional feature reconstruction on the images, and uses a softmax classifier to carry out classification; and an image fusion technology of the weighted guided filter and feature extraction are combinedthrough experimental simulation and verification, so that the target recognition precision is improved, the quantity of hidden layer neurons of the auto-encoder is greatly decreased and the calculation complexity is greatly reduced.

Description

technical field [0001] The invention relates to a target recognition algorithm based on a guided filter and an autoencoder, which can be applied to various military or civilian image processing systems. Background technique [0002] In modern high-tech warfare, the timely and accurate acquisition of battlefield information and the efficient assessment of the battlefield situation play a very important role in the struggle for military dominance on the battlefield. Synthetic Aperture Radar (SAR), as an important remote sensing imaging sensor, has a very wide range of applications in the fields of environmental monitoring, resource exploration and national defense and military affairs. Synthetic aperture radar has a certain ground and vegetation penetration ability, which is helpful to find man-made building targets such as airports, ports, bridges, roads, and military targets such as aircraft, tanks, and ships. SAR target recognition uses SAR image information to realize the...

AI Technical Summary

Problems solved by technology

In order to adapt to complex raw data, 64×64 neurons are required in the first layer of SAE, which leads to complex SAE structure and low network training efficiency

[0004] The recognition performance of neural networks requires a large amount of labeled data as training samples, but it is difficult to achieve in SAR objects, because complex SAR images usually require a complex network structure to fit them

As the complexity of the network structure increases, the training time of the network will be significantly extended

Images