Supercharge Your Innovation With Domain-Expert AI Agents!

Wave band self-adaptive hyperspectral image compression method based on 3D convolution auto-encoder

A convolutional self-encoding, hyperspectral image technology, applied in the field of hyperspectral image compression

Pending Publication Date: 2021-11-09
WUHAN UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

As an important method in deep learning, the 3D convolution kernel spectral dimension is not constrained by the input feature size, which makes it have the potential to solve the band adaptive compression problem.

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
  • Wave band self-adaptive hyperspectral image compression method based on 3D convolution auto-encoder
  • Wave band self-adaptive hyperspectral image compression method based on 3D convolution auto-encoder
  • Wave band self-adaptive hyperspectral image compression method based on 3D convolution auto-encoder

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] The specific compression process will be explained below in combination with examples and accompanying drawings. The specific compression steps of the hyperspectral image band adaptive compression method based on 3D convolutional autoencoder are as follows:

[0043] After the image tensor is compressed and processed by the encoder network, the hidden representation tensor is obtained, and then the hidden representation tensor is input into the quantizer for quantization processing to obtain the binary code stream for further compression, and finally the binary code stream is input to the decoder to obtain the reconstructed image. The training achieves network convergence and realizes rate-distortion optimization of images.

[0044] The encoder performs feature extraction on the input image to achieve preliminary compression. Such as figure 2 As shown, the encoder includes normalization, 3D convolution module 1, 3D convolution module 2, 3 3D residual modules, and 3D c...

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 technical scheme of the invention provides a wave band self-adaptive hyperspectral image compression method based on a 3D convolution auto-encoder. A network model is mainly divided into three modules, namely an encoder, a quantizer and a decoder. In consideration of the characteristic that the 3D convolution kernel spectrum dimension is not constrained by the input feature dimension, a 3D convolution auto-encoder is constructed, and convolution parameters are adjusted to ensure the invariance of the feature spectrum dimension in the feature extraction process, so that the high-performance compression and reconstruction of the hyperspectral image with any wave band number are realized, the method is of great significance in saving computing resources and promoting wide application of hyperspectral images.

Description

technical field [0001] The invention can be applied to the field of hyperspectral image compression, and realizes compression and reconstruction of hyperspectral images with different band numbers by using a 3D convolutional self-encoder framework. Background technique [0002] Compared with ordinary visible light images, hyperspectral images contain rich spectral information and are widely used in agriculture, remote sensing, medicine and other fields. With the development of spectral imaging technology, how to effectively solve the data transmission and storage pressure caused by the significant improvement of image spectral resolution and spatial resolution is an urgent problem to be solved in the application process of hyperspectral imagery. [0003] Since the hyperspectral images generated by different spectral imagers have different band numbers, designing a high-performance band-adaptive model for the difference in the band numbers of hyperspectral images to realize t...

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 Applications(China)
IPC IPC(8): G06T9/00G06N3/04G06N3/08G06T7/11
CPCG06T9/002G06T7/11G06N3/084G06T2207/20081G06T2207/20084G06T2207/20132G06T2207/10036G06N3/045
Inventor 种衍文陈林伟潘少明
Owner WUHAN UNIV
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