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

Multi-manifold-based multi-temporal hyperspectral image classification method

A technology of hyperspectral images and classification methods, which is applied in the field of multi-temporal hyperspectral remote sensing image classification, and can solve the problems of spectral drift of time-phase maps and difficulty in obtaining labels of multi-temporal hyperspectral data.

Active Publication Date: 2020-08-11
HARBIN INST OF TECH
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem of how to use the complementary information of multiple time phases to solve the difficulty of obtaining multi-temporal hyperspectral data labels, and the problem of obvious spectral drift between time-phase maps, and propose a multi-manifold-based multi-temporal hyperspectral Image Classification Methods

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
  • Multi-manifold-based multi-temporal hyperspectral image classification method
  • Multi-manifold-based multi-temporal hyperspectral image classification method
  • Multi-manifold-based multi-temporal hyperspectral image classification method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0024] Specific implementation mode one: combine figure 1 Describe this embodiment, the multi-manifold-based manifold alignment algorithm that maintains the global geometric structure of this embodiment, the specific process is:

[0025] Step 1. Input the spectral matrix X of samples in source image 1, source image 2 and target image s1 ,X s2 ,X t and their spatial coordinates L 1 , L 2 , L 3 , and X s1 ,X s2 The corresponding category label vector Y for each row 1 , Y 2 ;

[0026] Step 2. Calculate X s1 、X t The spatial spectral distance matrix d of 13 , X s2 、X t The spatial spectral distance matrix d of 23 , X s1 、X s2 The spatial spectral distance matrix d of 12 , for each type of sample in the source image, select k samples with the smallest spatial spectral distance in the target image to obtain three sets of data pairs that need to be matched;

[0027] Step 3. Calculate X s1 、X s2 、X t The geodesic distance matrix D s1,s1 、D s2,s2 、D t,t , and u...

specific Embodiment approach 2

[0032] Specific embodiment two: the difference between this embodiment and specific embodiment one is: the calculation of X in the second step s1 、X t The spatial spectral distance matrix d of 13 , X s2 、X t The spatial spectral distance matrix d of 23 , X s1 、X s2 The spatial spectral distance matrix d of 12 , for each type of sample in the source image, k samples with the smallest spatial spectral distance are selected in the target image, and three sets of data pairs that need to be matched are obtained. The specific process is:

[0033]

[0034]

[0035] d 12 = d 12 spectral d 12 spatial

[0036]

[0037]

[0038] d 13 = d 13 spectral d 13 spatial

[0039]

[0040]

[0041] d 23 = d 23 spectral d 23 spatial

[0042] in s1 is the source image 1, s2 is the source image 2, t represents the target image, x s1 i Denotes the ith sample of source image 1, x s2 i Denotes the ith sample of source image 2, x s2 j Denotes the jth sa...

specific Embodiment approach 3

[0046] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is: the calculation of X in the step three s1 、X s2 、X t The geodesic distance matrix D s1,s1 、D s2,s2 、D t,t , and using three sets of data pairs that need to be matched to obtain the distance matrix D of the spectra of different time phase diagrams s1,s2 、D s1,t 、D s2,t ; The specific process is:

[0047] Step 31. Calculate the respective geodesic distance D of the source image and the target image s1,s1 ,D s2,s2 ,D t,t ;

[0048] Step 32. Use three sets of data pairs that need to be matched to obtain the distance matrix D of the spectrum of the phase diagram at different times s1,s2 ,D s1,t ,D s2,t :

[0049]

[0050]

[0051]

[0052] Among them, a j is the first a in the matching data j total number of categories in the source image×k′; b j b in the matching data j , the total number of categories in the source image×k′; a j ,b j means x...

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

A multi-manifold-based multi-temporal hyperspectral image classification method, the invention relates to a multi-temporal hyperspectral remote sensing image classification method. The purpose of the present invention is to solve the problem of how to use complementary information of multiple time phases to solve the difficulty in obtaining multi-time phase hyperspectral data tags and the obvious spectral drift between time phase maps. The specific process is: 1. Input X s1 ,X s2 ,X t and their spatial coordinates L 1 , L 2 , L 3 , and Y 1 , Y 2 ; Two, calculate d 13 , d 23 , d 12 , each type of sample in the source image selects k samples with the smallest spatial spectral distance in the target image, and obtains three sets of data pairs that need to be matched; 3. Calculate D s1,s1 、D s2,s2 、D t,t , and D s1,s2 、D s1,t 、D s2,t ; 4. Adjust X s2 ,X t The data scale of the multi-manifold distance matrix D is constructed; 5. Get the projection f s1 , f s2 , f t 6. Obtain the classification label of the target phase. The invention is used in the field of image classification.

Description

technical field [0001] The invention relates to a multi-temporal hyperspectral remote sensing image classification method. Background technique [0002] With the development of optical sensors and spectroscopic technology, hyperspectral remote sensing imaging technology has developed rapidly on the basis of multispectral remote sensing imaging technology. Hyperspectral imaging technology combines imaging technology and spectral technology, and can simultaneously record the spatial structure information and rich spectral information of the area to be detected, and with the development of remote sensing technology, it can now provide large-area areas with spatial and temporal sequence Hyperspectral multi-temporal image data for ground object perception and monitoring. Multi-temporal hyperspectral images make multi-temporal analysis, multi-angle research, and accurate ground object change detection possible. However, it is unrealistic to provide sufficient label information f...

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): G06K9/00G06K9/62
CPCG06V20/194G06V20/13G06F18/24147
Inventor 谷延锋张美玲
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