SAR image classification method based on multi-feature and non-negative autoencoder

Abstract

The invention discloses a SAR image classification method based on multi-feature and non-negative automatic encoder, which belongs to the technical field of image processing. Extract the image block spatial domain features of SAR images based on the gray gradient co-occurrence matrix; extract the image block transform domain features of SAR images based on two-dimensional Gabor transform; combine the image block spatial domain features with the transform domain features; select SAR image block training Sample set and test sample set; use the training sample set to train the multi-layer non-negative autoencoder and softmax classifier; use the trained non-negative autoencoder network to classify; obtain the classification result map. The invention combines the spatial information and transform domain information of the SAR image, obtains the multi-dimensional features of the SAR image, and uses a non-negative autoencoder to optimize the features, which improves the distinguishability of the features, and then effectively improves the classification accuracy. It can be used for object classification and target recognition in high-resolution SAR images.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a SAR image classification method based on multi-features and non-negative automatic encoders in the field of object classification, which can be used for SAR image object classification and target recognition. Background technique [0002] Synthetic aperture radar (SAR) is an active imaging sensor with all-weather and all-time data acquisition capabilities, which has obvious advantages over traditional optical remote sensing technology. With the continuous development of remote sensing technology, the resolution of images acquired by SAR systems is getting higher and higher. High-resolution SAR images can reflect more detailed ground object information, which meets the needs of many practical applications. SAR image classification is an important content of SAR image interpretation, and has a wide range of applications in military reconnaissance, resource detection, geo...

AI Technical Summary

Problems solved by technology

The above method does not consider the influence of coherent speckle noise in SAR images and does not fully exploit the different features of SAR images, and does not effectively use the deep network to improve the distinction of features, thus affecting the classification effect

Images