Hyperspectral image classification method based on adaptive classification network model

Abstract

The invention discloses a hyperspectral image classification method and device based on an adaptive classification network model, electronic equipment and a storage medium, and the method comprises the steps: inputting a to-be-classified hyperspectral image into a pre-trained classification network model, and obtaining a classification result; wherein the classification network model is obtained by training a known training sample set; the training comprises a first training process and a second training process, in the first training process, a training network is a multi-layer stacked automatic encoder network, a first objective function adopted in the training process is a mean square error cross entropy loss function, similarity regularization constraint is introduced into the first objective function, and a pre-classification network model is obtained through training; in the second training process, the training network is a network in which an output layer of the pre-classification network model is replaced by a logistic regression layer, a second objective function adopted in the training process is a softmax cross entropy loss function, and a classification network model is obtained through training. The invention provides a hyperspectral image classification effect.

Description

technical field [0001] The invention belongs to the technical field of hyperspectral image processing, and in particular relates to a hyperspectral image classification method, device, electronic equipment and storage medium based on an adaptive classification network model. Background technique [0002] In remote sensing analysis, hyperspectral image classification is of great significance and is widely used in a variety of applications such as crop supervision, forest applications, urban development, and risk management. [0003] In recent years, image classification methods based on deep learning have been favored by researchers. Hyperspectral image classification is one of the research hotspots in the field of remote sensing. Many researchers build neural networks to automatically learn abstract features from hyperspectral remote sensing images, so as to continuously innovate ideal solutions to improve the accuracy of hyperspectral image classification. Common hyperspec...

AI Technical Summary

Problems solved by technology

[0005] However, the hyperspectral image classification method based on the adaptive classification network model proposed above still faces the problem of limited classification effect

Images