Hyperspectral image classification method based on semi-supervised WGAN-GP

Abstract

The invention discloses a hyperspectral image classification method based on semi-supervised WGAN-GP, which overcomes the problem that the existing technology is difficult to extract rich feature information under the condition of limited training data, and cannot be used to train the classifier using unlabeled samples, and the classification accuracy is low. The specific steps of the invention include: (1) inputting a hyperspectral image to be classified; (2) generating a sample set; (3) constructing semi-supervised WGAN-GP network; (4) training semi-supervised WGAN-GP network; (5) and classifying the test data. The invention can generate false hyperspectral data-assisted discriminator classification by the generator receiving noise in the semi-supervised WGAN-GP, and can fully utilize the limited sample to improve the classification precision, and can be used for hyperspectral images in the fields of fine agriculture and low quality research. Classify the target of the object.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a hyperspectral image classification based on semi-supervised Wasserstein distance and gradient penalty WGAN-GP (Wasserstein Generative Adversarial Net-Gradient Penalty) in the technical field of hyperspectral image classification method. The invention can be used to classify ground objects in hyperspectral images. Background technique [0002] Hyperspectral remote sensing images are satellite images captured by hyperspectral sensors, and each pixel has dozens or even hundreds of spectral bands. Therefore, it can provide rich information and has high spectral resolution, and can be widely used in many fields such as military affairs, agriculture, and environmental monitoring. The processing and analysis of hyperspectral images is extremely important in the field of international remote sensing, and hyperspectral image classification is an important research direc...

AI Technical Summary

Problems solved by technology

However, the disadvantage of this method is that the 3D convolutional neural network needs more training data to achieve the expected classification effect. When the amount of training data is limited, it is often difficult for the 3D convolutional neural network to extract effective features. Because of the classification of data, resulting in low classification accuracy

Although this method uses a deep convolutional network to extract features and obtain better classification results, the disadvantage of this method is that the built neural network is not assisted by other networks. It is difficult to extract rich features from small sample data, and the classification accuracy is low

Images