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

Hyperspectral remote sensing data classification method based on ensemble learning

A technology of hyperspectral remote sensing and integrated learning, which is applied in the field of spectral data classification, can solve the problem of low data classification accuracy, and achieve the effect of improving classification effect, high classification accuracy, and high data classification accuracy

Inactive Publication Date: 2014-09-03
HARBIN INST OF TECH
View PDF2 Cites 24 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the existing hyperspectral data classification method classifies data from the perspective of spectral dimension, and the data classification accuracy is low, and provides a hyperspectral remote sensing data classification method based on integrated learning

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 data classification method based on ensemble learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0030] Specific implementation mode one: the following combination figure 1 Describe this embodiment, the hyperspectral remote sensing data classification method based on integrated learning described in this embodiment, it comprises the following steps:

[0031] Step 1: Read the hyperspectral remote sensing data, use the principal component analysis method to calculate the eigenvalues ​​and eigenvectors of the hyperspectral remote sensing data, and then obtain the spectral characteristics of the hyperspectral remote sensing data; then use the gray level co-occurrence matrix to extract the high Spatial features of spectral remote sensing data;

[0032] Step 2: Integrate spectral features and spatial features into a multi-feature set;

[0033] Step 3: Determine the labeled samples from the multi-feature set and select training samples and test samples;

[0034] Step 4: Based on the integrated learning method, design the Adaboost integrated classification framework with featur...

specific Embodiment approach 2

[0037] Specific implementation mode two: this implementation mode further explains implementation mode one, and the method for obtaining the spectral features of the hyperspectral remote sensing data described in this implementation mode is:

[0038] First, calculate the covariance matrix of the hyperspectral remote sensing data, and use the principal component analysis method to obtain the eigenvalues ​​and eigenvectors of all the data according to the covariance matrix, and arrange the corresponding eigenvectors according to the order of the eigenvalues ​​from large to small, and Use the eigenvector as the weighting coefficient to calculate and obtain B principal component components, B is a positive integer; select the first N principal component components as the spectral features of the hyperspectral remote sensing data, and N is the value after rounding B / 2;

[0039] The method for obtaining the spatial characteristics of hyperspectral remote sensing data is as follows: e...

specific Embodiment approach 3

[0044] Specific implementation mode three: this implementation mode further explains implementation mode two, and the method for obtaining the multi-feature set in step two described in this implementation mode is:

[0045] Randomly select D numbers from 1 to N, D<N, extract the corresponding one-dimensional spectral features and two-dimensional spatial feature data according to the D number numbers in the spectral features and spatial features, and form a multi-dimensional matrix representation Feature set; repeat this process F-1 times to obtain F multi-feature sets.

[0046] Integrate spectral and spatial features into a multi-feature set. Such as figure 1 As shown in , in order to facilitate the subsequent integrated classification, F multi-feature sets need to be formed. Multi-feature sets include spectral features and spatial features. D digit numbers are expressed as {d 1 , d 2 ,...,d D}.

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 discloses a hyperspectral remote sensing data classification method based on ensemble learning, belonging to the technical field of spectral data classification and aiming at solving the problem of low data classification precision caused by the fact that an existing hyperspectral data classification method is used for classifying data from the aspect of spectral dimensions. The hyperspectral remote sensing data classification method comprises the following steps: firstly, reading hyperspectral remote sensing data to obtain spectral characteristics and spatial characteristics of hyperspectral remote sensing data; integrating the spectral characteristics and the spatial characteristics to obtain a multiple characteristic set; determining a marking sample and selecting training samples and testing samples according to the multiple characteristic set; designing an Adaboost integration and classification frame of characteristic difference based on an ensemble learning method, and training with the training sample to obtain F weak classifiers; and classifying the testing samples by the F weak classifiers. The hyperspectral remote sensing data classification method is used for classifying hyperspectral remote sensing data.

Description

technical field [0001] The invention relates to a hyperspectral remote sensing data classification method based on integrated learning, and belongs to the technical field of spectral data classification. Background technique [0002] With the development of remote sensing technology and the improvement of imaging equipment, hyperspectral remote sensing has become a new stage in optical remote sensing. Hyperspectral remote sensing can acquire many very narrow and spectrally continuous image data in the ultraviolet, visible, near-infrared and mid-infrared regions of the electromagnetic spectrum. This has made a new leap forward in people's ability to observe and detect surface features through remote sensing technology. In the hyperspectral data, the spatial structure and spectral characteristic information of the ground objects are included at the same time, which greatly improves the recognition and classification capabilities of the ground objects. However, the characteri...

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): G06K9/62G06K9/66G06V10/774G06V20/13
CPCG06V20/13G06V10/774
Inventor 陈雨时赵兴王强刘思宇
Owner HARBIN INST OF TECH
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