Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Hyper-spectral image classification method based on singular value decomposition and neighborhood space information

A singular value decomposition and spatial information technology, applied in the field of hyperspectral image classification, can solve the problems of consuming a lot of time and money, having a large impact on classification results, and consuming a large amount of training time, so as to improve classification accuracy, improve automation level, shorten The effect of classification time

Inactive Publication Date: 2016-07-06
XIAMEN UNIV OF TECH
View PDF4 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Sample acquisition often takes a lot of time and money
The support vector machine method can also obtain high classification accuracy in the case of small samples, but the support vector machine needs to select a suitable kernel function and its corresponding parameters. Different parameters often have a greater impact on the classification results, and through cross-validation It often takes a lot of training time to obtain suitable parameters by other means.

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

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] The technical solutions and beneficial effects of the present invention will be described in detail below in conjunction with specific embodiments.

[0016] The present invention provides a hyperspectral image classification method based on singular value decomposition and neighborhood spatial information, comprising the following steps:

[0017] Step 1, input data: input the m×n training sample matrix of each class Among them, m is the number of training samples, n is the dimension of the hyperspectral vector, j represents its corresponding category (such as waters, roads, trees, etc.); at the same time, input the l×n test sample matrix D l×n , l is the number of test samples.

[0018] The following will enter the iterative loop to find the most suitable r value to get the best classification scheme.

[0019] Step 2, using formula A m×n =U m×r S r×r V r×n For the training sample matrix Perform singular value decomposition, where U m×r and V r×n are left 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 discloses a hyper-spectral image classification method based on singular value decomposition and neighborhood space information, comprising the following steps: inputting a training sample matrix of each category, and carrying out singular value decomposition on the training sample matrixes to get a right singular matrix corresponding to jth-category training samples; for data in the training sample matrixes, using a least squares method to calculate the residual of the data corresponding to each category, comparing the residuals of the data corresponding to all the categories, and classifying the data to the category corresponding to the smallest residual; and repeating the steps to get the category of each training sample, comparing the category of each training sample with the original category, getting a parameter making the classification accuracy rate of the training samples highest through iterative comparison, classifying each data in a test sample matrix, and outputting a classification result matrix. Through the classification method, the precision of classification is improved, the time for classification is shortened, and the level of automation of hyper-spectral image classification is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a hyperspectral image classification method based on singular value decomposition and neighborhood space information. Background technique [0002] Most hyperspectral image supervised classification methods need enough training samples to support the classification results, such as linear discriminant analysis, naive Bayesian, neural network, etc. Sample acquisition is often time-consuming and expensive. The support vector machine method can also obtain high classification accuracy in the case of small samples, but the support vector machine needs to select a suitable kernel function and its corresponding parameters. Different parameters often have a greater impact on the classification results, and through cross-validation Obtaining appropriate parameters by other means often consumes a lot of training time. Contents of the invention [0003] The purpose of the pres...

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): G06K9/62
CPCG06F18/2415
Inventor 廖文超张斌朱述龙
Owner XIAMEN UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products