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

Hyperspectral remote sensing data classification method based on deep learning

A hyperspectral remote sensing and deep learning technology, applied in the field of hyperspectral data classification, can solve the problem of low classification accuracy, achieve accurate data classification and reduce the effect of impact

Active Publication Date: 2017-05-03
HARBIN INST OF TECH
View PDF3 Cites 0 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 of low classification accuracy in existing methods for classifying hyperspectral data containing nonlinear features, and to provide a method for classifying hyperspectral remote sensing data based on deep 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 deep learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0033] Specific embodiment one: the following combination figure 1 Describe this embodiment, the deep learning-based hyperspectral remote sensing data classification method described in this embodiment, which includes the following steps:

[0034] Step 1: Read the hyperspectral raw data, use the principal component analysis method to obtain the eigenvalues ​​and eigenvectors of the hyperspectral raw data, and then obtain the spectral eigenvectors of the hyperspectral raw data; Domain information extraction to obtain spatial feature information of hyperspectral raw data;

[0035] Step 2: Integrate the spectral feature vector and spatial feature information of the original hyperspectral data to obtain integrated hyperspectral data;

[0036] Step 3: Determine the labeled samples from the hyperspectral integration data, and select training samples and test samples from the labeled samples;

[0037] Step 4: Based on the deep learning method, use the training samples to pre-train ...

specific Embodiment approach 2

[0040] Embodiment 2: This embodiment further describes Embodiment 1, and the method for obtaining the spectral feature vector of the hyperspectral raw data is as follows:

[0041] First, the covariance matrix of the hyperspectral raw data is calculated, and the eigenvalues ​​and eigenvectors of all hyperspectral raw data are calculated and obtained by principal component analysis according to the covariance matrix, and the corresponding features are arranged in the order of eigenvalues ​​from large to small. vector; use the eigenvectors as the weighting coefficient to obtain the B-band principal component components of all eigenvectors, and use all the eigenvectors containing the B-band principal component components as the spectral eigenvectors of the hyperspectral raw data; B is a positive integer;

[0042] The method of obtaining the spatial feature information of the hyperspectral raw data is as follows:

[0043] Select the first N principal component components in the B b...

specific Embodiment approach 3

[0047] Embodiment 3: This embodiment further describes Embodiment 2, and the specific method for obtaining hyperspectral integration data is: i , y i ) spectral information is regarded as a spectral feature vector of length B, and all the pixel points (x i , y i ), a spectral eigenvector of length B and a length of N×w 2 The spatial feature information of , integrated into a length of (B+N×w 2 ) vector, as the hyperspectral integration data.

[0048] In this embodiment, in the hyperspectral data after information integration, each pixel will have a length of (B+N×w containing spectral information and spatial information) 2 ) feature vector.

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 deep learning, and belongs to the technical field of hyperspectral data classification. The invention aims to solve a problem of low classification precision of a method for classifying hyperspectral remote sensing data with nonlinear characteristics. The hyperspectral remote sensing data classification method comprises the following steps: firstly, processing hyperspectral original data to obtain the spectral feature vector and the spatial feature information of the hyperspectral original data; then, integrating the spectral feature vector with the spatial feature information; confirming labeled samples by hyperspectral integrated data, selecting a training sample and a test sample from the labeled samples; Pre-training a multi-layer restricted Boltzmann machine which forms a deep network by the training sample; carrying out supervised learning to the network formed by the multi-layer restricted Boltzmann machine through the training sample; and inputting the test sample into the trimmed network formed by the multi-layer restricted Boltzmann machine to realize hyperspectral remote sensing data classification. The invention is used for the hyperspectral remote sensing data classification.

Description

technical field [0001] The invention relates to a hyperspectral remote sensing data classification method based on deep learning, and belongs to the technical field of hyperspectral data classification. Background technique [0002] With the successful development of airborne and spaceborne hyperspectral sensors, the spectral resolution of remote sensing data has been greatly improved, which has solved the problems that multispectral remote sensing cannot solve. At the same time, the processing of high-dimensional data has also has become an urgent problem. The high dimension of the data comes from the high spectral resolution. For the hyperspectral remote sensing with spectral resolution reaching the nanometer level, the data dimension on the spectrum can reach hundreds of dimensions. The high-dimensional space of the data is very limited. [0003] At present, the mainstream method to solve the high-dimensional problem of hyperspectral data is feature extraction. However...

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): G06K9/66
Inventor 陈雨时赵兴王强时春雨
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