A Tensor Representation Based Target Detection Method for Polarized Hyperspectral Images

A polarization hyperspectral and target detection technology, which is applied in image analysis, image enhancement, image data processing, etc., can solve the problems of low information utilization rate and low accuracy of polarization hyperspectral image target detection, and achieve accurate target detection, reduce The effect of simplifying the calculation process and improving the accuracy

Active Publication Date: 2018-12-11
黑龙江省工业技术研究院
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the problems of low detection accuracy and low information utilization rate of existing polarization hyperspectral image targets, and propose a polarization hyperspectral image target detection method based on tensor representation

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 Tensor Representation Based Target Detection Method for Polarized Hyperspectral Images
  • A Tensor Representation Based Target Detection Method for Polarized Hyperspectral Images
  • A Tensor Representation Based Target Detection Method for Polarized Hyperspectral Images

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0053] Specific Embodiment 1: A tensor representation-based polarization hyperspectral image target detection method in this embodiment is specifically carried out in accordance with the following steps:

[0054] The concept of tensor was first proposed by Woldemar Voigt in the late 1790s. Later, Ricci and his students gradually perfected the concept of tensor and related operations in their research. From an algebraic point of view, the tensor can be regarded as a further extension of the matrix. The rows of the matrix are modulo-1 fibers, and the columns of the matrix are modulo-2 fibers. The third-order tensor has more tube fibers, such as Figure 1a , 1b , 1c, and 1d show the format of a third-order tensor. For the vector space U (1) ,U (2), ...,U (M) , defining their outer product space for tensors.

[0055] For the basic operations on tensors involved in the follow-up, this section makes a brief introduction. The first is the n-module expansion of the tensor, for...

specific Embodiment approach 2

[0078] Specific embodiment two: the difference between this embodiment and specific embodiment one is: in said step two, on the basis of the data model of the polarized hyperspectral image under the tensor representation, the fourth-order tensor matching detection based on Tucker decomposition is obtained The model of the algorithm (FTMF) judges the detection result according to the model of the fourth-order tensor matching detection algorithm (FTMF) based on Tucker decomposition; the specific process is called:

[0079] Tucker decomposition is used for the training samples to extract the information of the target spectral dimension and polarization dimension, and the first principal component (i.e. the first column) of the factor matrix of the target spectral dimension and polarization dimension is selected to replace the target spectrum and polarization spectrum, and then according to the spectral dimension and polarization Spectral matched filter and polarization matched fil...

specific Embodiment approach 3

[0081] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is: the spectral matching filter and polarization matching filter are constructed according to the information of the spectral dimension and the polarization dimension, and the fourth-order tensor matching detection based on Tucker decomposition is obtained The model of the algorithm (FTMF) judges the detection result according to the model of the fourth-order tensor matching detection algorithm (FTMF); the specific process is called:

[0082] Spectral Matched Filter H 3 Constructed as follows:

[0083]

[0084] Among them, S 3 Indicates the target spectrum, R 3 =E(Y (3) Y (3) T ) is the data to be detected Y (3) The covariance matrix of , Represents a fourth-order tensor The 3-mode expansion, In is the size of the nth dimension of the image, the value of n is 1, 2, 3 or 4, and T is the transpose;

[0085] will pair the spectrally matched filter H3 and 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

The invention discloses a tensor expression-based polarized hyperspectral image target detection method, relates to a polarized hyperspectral image target detection method, and aims at solving the problem that the conventional polarized hyperspectral image target detection is low in correctness and low in information utilization rate. The method comprises the following steps of: 1, establishing a data model of a polarized hyperspectral image under tensor expression: expressing the polarized hyperspectral image into a four-order tensor so as to obtain the data model of the polarized hyperspectral image; and 2, obtaining a model of a Tucker decomposition-based four-order tensor matching detection algorithm on the basis of the data model of the polarized hyperspectral image under tensor expression, and judging a detection result according to the model of the Tucker decomposition-based four-order tensor matching detection algorithm. The method disclosed by the invention is used for polarized hyperspectral image target detection.

Description

technical field [0001] The invention relates to a method for detecting a polarized hyperspectral image target. Background technique [0002] The Stokes vector more completely represents the polarization characteristics of light, and more accurately describes the composition and surface characteristics of the object. However, for polarized hyperspectral images, which have multi-dimensional information of polarization, intensity, spectrum and space, it is difficult for the Stokes vector to completely describe them, and it ignores the polarization dimension information of polarized hyperspectral images that changes with the polarization angle; and the traditional The current polarization hyperspectral image detection algorithm also ignores the spatial information of the image, and it is difficult to realize the common use of multi-dimensional information of polarization hyperspectral images. In addition, when the data volume of the polarization hyperspectral image is large, th...

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): G06T7/33
CPCG06T2207/10036
Inventor 张钧萍谭建
Owner 黑龙江省工业技术研究院
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
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap