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

A hyperspectral video image gas detection method based on cumulative tensor decomposition

A video image and gas detection technology, which is applied in the field of hyperspectral video image gas detection, can solve the problems of not realizing the effective use of time phase information of hyperspectral video data, low detection accuracy, etc., to achieve improved accuracy, wide application scenarios, good The effect of detection performance

Active Publication Date: 2021-02-02
HARBIN INST OF TECH
View PDF6 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 that the time phase information of hyperspectral video data is not effectively utilized and the detection accuracy is low in the existing hyperspectral video image gas detection method, and a hyperspectral method based on cumulative tensor decomposition is proposed. Video image gas detection 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 hyperspectral video image gas detection method based on cumulative tensor decomposition
  • A hyperspectral video image gas detection method based on cumulative tensor decomposition
  • A hyperspectral video image gas detection method based on cumulative tensor decomposition

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0037] Specific Embodiment 1: In this embodiment, a hyperspectral video image gas detection method based on cumulative tensor decomposition, the specific process is as follows:

[0038] Step 1: Express the hyperspectral video image sequence composed of the previous t moments as a fourth-order cumulative tensor Will As the initial fourth-order cumulative tensor Calculated using the classic alternating least squares method As a result of its CP decomposition, we get factor matrix in four dimensions i represents the dimension, 1 represents the space-X dimension, 2 represents the space-Y dimension, 3 represents the spectral dimension, and 4 represents the time dimension;

[0039] based on and Initialize two sets of auxiliary matrices on the three non-time dimensions corresponding to time t and n=1,2,3, n represents dimension, 1 represents space-X dimension, 2 represents space-Y dimension, 3 represents spectral dimension;

[0040] Step 2: For a newly added frame ...

specific Embodiment approach 2

[0046] Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is that in the first step, the hyperspectral video image sequence composed of the previous t moments is expressed as a fourth-order cumulative tensor Will As the initial fourth-order cumulative tensor Calculated using the classic alternating least squares method As a result of its CP decomposition, we get factor matrix in four dimensions based on and Initialize two sets of auxiliary matrices on the three non-time dimensions corresponding to time t and The specific process is:

[0047] Given a hyperspectral video image sequence, the hyperspectral video image sequence composed of the first t moments is expressed as a fourth-order cumulative tensor Will As the initial fourth-order cumulative tensor Using the classic Alternating Least Squares algorithm (Alternating Least Squares, ALS) to get factor matrix in four dimensions

[0048] Among them, the first dimen...

specific Embodiment approach 3

[0058] Embodiment 3: The difference between this embodiment and Embodiment 1 or 2 is that in the step 2, for a newly added frame of image at time t+1, it itself can be expressed as a three-dimensional hyperspectral image Will Append the rank 4 cumulative tensor initial in step 1 The time dimension of , such that the initial fourth-order cumulative tensor Expand in the time dimension to obtain a new fourth-order cumulative tensor corresponding to time t+1 The specific process is:

[0059] Newly added 3D hyperspectral image stereo at time t+1 Represented as a third-order tensor, the initial fourth-order cumulative tensor Added to the time dimension of Get the new fourth-order cumulative tensor corresponding to time t+1

[0060] in, The superscript of represents the order of the tensor, and the subscript represents the moment of the video sequence corresponding to a single third-order tensor; Append the rank 4 cumulative tensor initial in step 1 The time dim...

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 hyperspectral video image gas detection method based on cumulative tensor decomposition, the invention relates to a hyperspectral video image gas detection method. The purpose of the present invention is to solve the problem of low detection accuracy in the existing hyperspectral video image gas detection method. The process is: Think of the video as an expanding fourth-order cumulative tensor. On the basis of CP decomposition of the initial cumulative tensor, for each new image at each moment, a new cumulative tensor is added, updated on the result of CP decomposition of the cumulative tensor at the previous moment, and the updated factor The matrix is ​​used to approximate the new image, and the threshold judgment is made on the fitting degree of the new image and its CP approximation result. When it is less than a given threshold, it is considered that there is no gas target in the new image, and the initial cumulative tensor and factor matrix are updated; otherwise, it is considered that there is a gas target in the new image, and its CP approximate residual tensor is obtained, and the residual The maximum value of the difference tensor in the spectral dimension is used as the detection result.

Description

technical field [0001] The invention relates to a hyperspectral video image gas detection method. Background technique [0002] With the continuous progress of human society, many gas pollutants harmful to the environment and human body will be produced in the process of industrial production, but many gases are colorless and odorless, and usually diffuse rapidly over time, which makes ordinary imaging It is difficult for the sensor to detect it effectively. Many gases have their unique spectral characteristics in the thermal infrared band. We can record the diffusion process of gases based on infrared hyperspectral video cameras, and use the collected infrared hyperspectral video data to detect them. The study of gas technology has great theoretical significance and practical application value. On the civilian side, we can monitor emissions from industrial sources in real time, minimizing their impact on the environment and their potential harm to populations living nearb...

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): G06K9/00G01N21/3504
CPCG01N21/3504G06V20/41G06V20/48
Inventor 谷延锋谭苏灵
Owner HARBIN INST OF TECH
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com