Unlock instant, AI-driven research and patent intelligence for your innovation.

A High-Resolution Multi-temporal Remote Sensing Image Classification Method Based on Multi-Connection Decision Manifold Alignment

A high-resolution, remote-sensing image technology, which is applied in the directions of instruments, calculations, character and pattern recognition, etc., can solve the problems of spectral aliasing, few spectral bands, and low accuracy of multi-time relative alignment classification, so as to expand the application scope and improve the capabilities, to achieve the effect of deepening utilization

Active Publication Date: 2019-06-11
天岸马科技(黑龙江)有限公司
View PDF7 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 multi-temporal alignment classification under high-scoring conditions, that is, to solve the problem of low accuracy of traditional multi-temporal alignment classification caused by fewer spectral bands and serious spectral aliasing, and propose a multi-temporal alignment based on multi-connection decision manifold Quasi-high-resolution multi-temporal remote sensing image classification method

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
  • A High-Resolution Multi-temporal Remote Sensing Image Classification Method Based on Multi-Connection Decision Manifold Alignment
  • A High-Resolution Multi-temporal Remote Sensing Image Classification Method Based on Multi-Connection Decision Manifold Alignment
  • A High-Resolution Multi-temporal Remote Sensing Image Classification Method Based on Multi-Connection Decision Manifold Alignment

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0029] Specific implementation mode one: combine figure 1 To illustrate this embodiment, the specific process of a high-resolution multi-temporal remote sensing image classification method based on multi-connection decision manifold alignment in this embodiment is as follows:

[0030] Step 1. Input the spectral matrices A and B of all spatial points in the source phase and the target phase, and the corresponding category label vector Y of each row in A;

[0031] Step 2. Calculate the similarity matrix W between A, B and each row in A and B respectively 1 , W 2 and W 12 ;

[0032] Step 3. Calculate W through multi-connection decision-making 1 , W 2 The corresponding source-time-phase single-connection Laplacian graph G 1 , target phase multi-connected Laplacian graph G 2 , and the source-phase and target-phase single-connection similarity matrix W 12 Multi-connection decision-making optimization;

[0033] Step 4, put A, B, G 1 , G 2 and W 12 Input into the independe...

specific Embodiment approach 2

[0035] Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is: in the step 2, calculate the similarity matrix W between A, B and each row in A and B 1 , W 2 and W 12 ; The specific process is:

[0036]

[0037]

[0038]

[0039] In the formula, W respectively 1 , W 2 and W 12The elements in row i and column j of , σ is the control parameter of similarity; A i,: For all elements of the i-th row of the matrix A, A j,: For all elements of the jth row of the matrix A, B i,: For all elements of the i-th row of the matrix B, B j,: is all the elements of the jth row of the matrix B, and : is all the columns of the ith row.

[0040] Other steps and parameters are the same as those in Embodiment 1.

specific Embodiment approach 3

[0041] Specific implementation mode three: combination Figure 2a , Figure 2b , Figure 2c Describe this implementation mode. The difference between this implementation mode and the specific implementation mode 1 or 2 is that in the step 3, W is calculated by means of multi-connection decision-making. 1 , W 2 The corresponding source-time-phase single-connection Laplacian graph G 1 , target phase multi-connected Laplacian graph G 2 , and the source-phase and target-phase single-connection similarity matrix W 12 Multi-connection decision-making optimization; the specific process is:

[0042] Step 31, in turn for W 1 Each row of performs the following operations: find W 1 The first p values ​​sorted from large to small in the i-th row, and find the corresponding label of the position of the p value in the label vector Y, count the label category with the most occurrences among the p labels that appear, and then find The maximum position t of the similarity value of 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

A high-resolution multi-temporal remote sensing image classification method based on multi-connection decision manifold alignment, the invention relates to a high-resolution multi-temporal remote sensing image classification method. The purpose of the present invention is to solve the problem of multi-temporal alignment classification under high-scoring conditions. The specific process is: 1. Input the spectral matrices A and B of all spatial points in the source phase and the target phase, and the corresponding category label vector Y of each row in A; 2. Calculate the difference between A, B and each row in A and B respectively W 1 , W 2 and W 12 ; 3. Calculate W through multi-connection decision-making 1 , W 2 Corresponding G 1 , G 2 , and W 12 The multi-connection decision-making optimization of ; four, A, B, G 1 , G 2 and W 12 Input into the independent spectral alignment model to obtain the mapping matrix F of A and B in the alignment space 1 and F 2 ; 5. Classify through the KNN classification model to 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 high-resolution multi-temporal remote sensing image classification method. Background technique [0002] Multispectral optical remote sensing is an important branch in the field of remote sensing. The current multispectral optical remote sensing technology is gradually developing from low spatial resolution to high spatial resolution, from typical MODIS data with a resolution of 1000 meters to Landsat with a spatial resolution of 30 meters. The data is in my country's Gaofen No. 1 data with a resolution of 8 meters and Gaofen No. 2 data with a resolution of 3.2 meters. High spatial resolution multi-spectral remote sensing images not only contain the spectral information of ordinary ground object images, but also contain the geometric structure and texture information of ground objects. The application of high-resolution remote sensing image technology not only improves the traditional analysis scale, but also expands the trad...

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): G06K9/62
CPCG06F18/24147
Inventor 谷延锋高国明
Owner 天岸马科技(黑龙江)有限公司