Attention and generative adversarial network-based optical remote sensing image retrieval method

Abstract

The invention discloses an attention and generative adversarial network-based optical remote sensing image retrieval method, and mainly solves the problem of low optical remote sensing image retrievalprecision in the prior art. The method comprises the following specific steps: (1) constructing a deep convolutional network; (2) constructing an attention network; (3) constructing a generative adversarial network; (4) constructing a hash learning network; (5) training a network; (6) acquiring a hash coding vector of each optical remote sensing image; and (7) retrieving an optical remote sensingimage. According to the method, the attention network is constructed, discriminable features of the image are extracted, and the expression ability of image features is improved; a generative adversarial network is constructed, an image hash coding vector is extracted, and quantization errors are reduced; and finally, the retrieval precision of the optical remote sensing image is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to an optical remote sensing image retrieval method based on attention and generation confrontation network in the technical field of optical remote sensing image retrieval. The invention can quickly and accurately search for the image that the user is interested in from a large amount of optical remote sensing images. Background technique [0002] With the development of satellite remote sensing and aerial remote sensing technology, the data volume and image resolution of remote sensing images are increasing, and more useful data and information can be obtained from remote sensing images. For the application of different occasions, the processing of remote sensing images also has different requirements, so in order to effectively analyze and manage these remote sensing image data, it is necessary to quickly query and retrieve interesting images from the massive remot...

AI Technical Summary

Problems solved by technology

However, the disadvantage of this method is that only relying on the bag-of-words model to encode the features of the image to obtain the image feature vector, using this image feature vector to retrieve a large number of optical remote sensing images, the time consumption is very large

However, this method still has shortcomings: since the hash function is discrete, there is a quantization error from the deep feature of the image to the hash code. However, this method does not have an effective mechanism to reduce the quantization error, which affects the final search precision

Images