Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Hyperspectral remote sensing image classification method based on semi-supervised sparse discriminative embedding

A sparse discriminative embedding, hyperspectral remote sensing technology, applied in character and pattern recognition, instrumentation, computing, etc., can solve the problem of not effectively utilizing discriminative information

Inactive Publication Date: 2016-08-17
CHONGQING UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the SPP algorithm does not need to mark the training samples, it does not effectively use the discriminative information provided in the marked samples.

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 based on semi-supervised sparse discriminative embedding
  • Hyperspectral remote sensing image classification method based on semi-supervised sparse discriminative embedding
  • Hyperspectral remote sensing image classification method based on semi-supervised sparse discriminative embedding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0136] The Washington DC Mall hyperspectral remote sensing image dataset used in this experiment is a local area of ​​the National Mall in Washington DC. The dataset has 191 bands with a band spacing of 10nm and a spatial resolution of 3m. The known categories of ground features include seven categories: "building", "forest", "(stone path) path", "road", "lawn", "lake" and "shadow". In the experiment, four classification experiments were carried out for each classification method involved in the comparison. The four classification experiments randomly selected 2, 4, 6, and 8 data points from each type of ground object as classified points, and obtained from the remaining training samples. Randomly select 60 unlabeled data as unlabeled samples to form a training sample set, and all the remaining data points are used as test samples. The four classification experiments are recorded as 2-lab, 4-lab, 6-lab, 8-lab, respectively. lab, for each classification method involved in the ...

Embodiment 2

[0141] The Indian Pine hyperspectral remote sensing image dataset used in the experiment covers an agricultural area in the northwest of Indiana, USA. The data point space size of this dataset is 145×145, and there are 220 bands and 17 known object categories. Considering the influence of noise, this experiment selects 200 bands and selects 6 categories from the ground object categories with more data points for experiments. The fifth categories of the 6 categories are "Hay_windrowed", "Soybeans_min", "Woods", "Corn_notill", "Grass_pasture", "Grass_trees". In the experiment, four kinds of classification experiments were carried out for each classification method participating in the comparison. The four classification experiments randomly selected 2, 4, 6, and 8 data points with category labels from each type of ground object, and selected data points from the remaining training samples. Randomly select 60 unlabeled data as unlabeled samples to form a training sample set, and...

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 provides a hyperspectral remote sensing image classification method based on semi-supervised sparse discriminative embedding, which uses semi-supervised sparse discriminative embedding algorithm to reduce the dimension of hyperspectral remote sensing images, combines the advantages of the nearest neighbor manifold structure and sparsity, not only The sparse reconstruction relationship between samples is preserved, and the natural discriminative ability of sparse representation is used, without the need to artificially select the value of the neighbor parameter, which alleviates the difficulty of selecting the neighbor parameter to a certain extent. At the same time, a small number of labeled training samples and some unmarked training samples are used To discover the intrinsic properties and low-dimensional manifold structure hidden in high-dimensional data, it can improve the classification accuracy of ground object categories in hyperspectral remote sensing images; at the same time, the method of the present invention treats marked data and unlabeled data differently, Maximize the clusterability between data points of the same ground object category, so as to help improve the classification accuracy of the ground object category in hyperspectral remote sensing images on the other hand.

Description

technical field [0001] The invention relates to the technical field of hyperspectral data processing methods and applications, in particular to a hyperspectral remote sensing image classification method based on semi-supervised sparse discriminant embedding. Background technique [0002] Hyperspectral remote sensing technology has developed rapidly since the 1980s. Its images record the continuous spectrum of ground objects, contain richer information, and have the ability to identify more types of ground objects and classify objects with higher accuracy. However, since hyperspectral data consists of a large number of bands to form a high-dimensional feature space, the complexity of most algorithms increases exponentially with the number of dimensions, and its processing requires a greater amount of calculation, and its bands are highly correlated and redundant. , at the same time, there are problems such as high dimensionality, easy to obtain ideal results due to Hughes phe...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62G06K9/66
Inventor 黄鸿曲焕鹏
Owner CHONGQING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com