Synthetic aperture radar remote sensing image ocean floating raft recognition method for deep-cooperative sparse coding network

Abstract

The present invention provides a synthetic aperture radar remote sensing image ocean floating raft recognition method for a deep-cooperative sparse coding network, and belongs to the field of computerremote sensing image processing and pattern recognition. The method comprises the following steps: preprocessing a synthetic aperture radar remote sensing image to obtain a remote sensing image withan accurate geographic coordinate position, low speckle noise and high image visibility; extracting texture features and contour features of the radar remote sensing image; performing super-pixel segmentation on the texture image and the contour image; and finally, inputting a super-pixel block feature image into a classification recognizer to obtain a classification recognition result. The effectand benefit of the method provided by the present invention is that: the supervised classification accuracy of the synthetic aperture radar remote sensing image can be greatly improved.

Description

technical field [0001] The invention belongs to the field of computer synthetic aperture radar (SAR) remote sensing image processing and pattern recognition, and relates to a method for identifying marine floating rafts in synthetic aperture radar remote sensing images using a deep collaborative sparse coding network. Background technique [0002] The traditional clustering method only has a good clustering effect on the target sample data of spherical distribution. However, in the actual SAR image processing process, it is usually impossible to know the distribution type of floating raft culture in advance, so a supervised algorithm is used. [0003] According to the characteristics of SAR image remote sensing data, the supervised algorithm learns through training samples, so as to realize the classification of different types of targets. Zhang Hong et al proposed object-oriented high-resolution SAR image processing for marine target recognition and extraction (Zhang Hong, ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the visible light remote sensing image in the existing technology cannot completely and accurately acquire the breeding target, and the marine synthetic aperture radar remote sensing image contains a large amount of coherent noise, the image feature is single, and the target recognition is difficult

Images