Hyperspectral remote sensing image classification method

A technology of hyperspectral remote sensing and image classification, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as classification of difficult hyperspectral remote sensing images, improve classification performance and generalization ability, and expand data sets Effect

Active Publication Date: 2021-11-26
ANHUI UNIVERSITY
View PDF9 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to solve the defect that it is difficult to classify hyperspectral remote sensing images in the prior art, and provide a hyperspectral remote sensing image classification method to solve the above problems

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Hyperspectral remote sensing image classification method
  • Hyperspectral remote sensing image classification method
  • Hyperspectral remote sensing image classification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0154] In order to verify the influence of different parameter settings in the method proposed by the present invention on the classification accuracy, several factors that affect the classification effect of the model in the MSRA-G method will be analyzed, mainly including dimensionality reduction d, input sample window size w, and learning rate lr. The batch size is unified to 64, the iteration is 200 times, and the average value of the classification accuracy of 10 experiments is used as the experimental result.

[0155] Figure 5 It shows the impact of the classification method MSRA-G on the classification accuracy under different dimensions d in the present invention. from Figure 5 It can be observed that different dimensionality reduction d makes OA perform differently, basically showing a trend of first increasing and then decreasing. For the IP dataset, when d is 14, OA reaches the highest value, and then the OA value is basically stable. While the UP and SA data ...

Embodiment 2

[0158] Embodiment 2: In order to further verify the performance of the algorithm in this paper, this embodiment uses three data sets of IP, UP, and SV for verification, and randomly selects 5%, 1%, and 0.5% from each type of ground object as a training sample set, The rest are used as the test sample set. And use REF-SVM, 3D-CNN, MSDN, HybridSN, SSRN and R-HybridSN six hyperspectral image classification methods as the comparative experimental objects, the classification results take the mean of ten experimental results, and record the standard deviation, so as to verify MSRA - Classification performance of the G method.

[0159] Table 2 Comparison table of classification accuracy of different classification methods on IP dataset

[0160]

[0161]

[0162] right figure 2 The IP hyperspectral images shown are classified, and the MSRA-G classification method proposed in the present invention uses GANs to generate synthetic samples to achieve the purpose of expanding the ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a hyperspectral remote sensing image classification method. Compared with the prior art, the problems of insufficient hyperspectral remote sensing image feature extraction and non-ideal classification precision under limited sample size are solved. The method comprises the following steps: obtaining and preprocessing a hyperspectral remote sensing image; constructing and training a generative adversarial network; obtaining an expansion training sample; constructing a multi-scale residual attention network; training a multi-scale residual attention network; obtaining a hyperspectral remote sensing image to be classified; and obtaining a hyperspectral remote sensing image classification result. According to the method, an ideal hyperspectral remote sensing image classification result can be obtained under the condition of insufficient training samples.

Description

technical field [0001] The invention relates to the technical field of hyperspectral remote sensing images, in particular to a hyperspectral remote sensing image classification method. Background technique [0002] As an important means of earth observation, hyperspectral remote sensing can obtain fine ground object attribute information, and has received enough attention in recent years. Due to the rapid development of satellite sensor technology, a large number of hyperspectral remote sensing images have been captured, which have rich spectral information and spatial information, which brings new opportunities for the application of hyperspectral remote sensing technology. Hyperspectral classification aims to train a classifier based on some labeled pixel samples, and then predict the labels corresponding to other pixel samples in the image to obtain the spatial distribution of different objects in the image. However, in the process of hyperspectral image classification, ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06N3/048G06N3/045G06F18/2135G06F18/2415
Inventor 赵晋陵胡磊黄林生梁栋徐超黄文江
Owner ANHUI UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap