RGBD image joint recovery method based on double-flow network

An RGB image, joint restoration technology, applied in image enhancement, image analysis, image data processing and other directions, can solve the problem of missing noise in depth images, and achieve the effect of low time complexity and convenient application

Active Publication Date: 2020-05-05
SOUTH CHINA UNIV OF TECH
View PDF1 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problems existing in the prior art that the processing of RGBD image data is based on the fact that only the depth

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
  • RGBD image joint recovery method based on double-flow network
  • RGBD image joint recovery method based on double-flow network
  • RGBD image joint recovery method based on double-flow network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] The present invention will be further described below in conjunction with drawings and embodiments.

[0058] see Figure 1-2 , a kind of RGBD image joint restoration method based on two-stream network, comprising:

[0059] S1, acquiring an RGBD image database for training and testing; the RGBD image database is a large-scale RGBD image dataset KITTI depth competition dataset in an unmanned driving scene.

[0060] S2, the RGBD image database is divided into a training data set and a test data set, and the RGBD image of the RGBD image database is preprocessed as a network input; in this embodiment, step S2 includes: the RGBD image database is divided into a training data set And the test data set, the depth image and the corresponding RGB image in the training data set are taken as a group; the RGB image and the depth image of each group are cut to an appropriate size, and normalized.

[0061] S3, train the dual-stream convolutional network model according to the traini...

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 relates to an RGBD image joint recovery method based on a double-flow network. The method comprises the steps of: S1, obtaining an RGBD image database used for training and testing; S2,dividing the RGBD image database into a training data set and a test data set, and preprocessing RGBD images of the RGBD image database; S3, training a double-flow convolutional network model according to the training data set, and storing the trained network parameters; and S4, inputting the test data set into the double-flow convolutional network model for joint recovery, and testing the recovery degree. According to the method, the degraded RGB image and the corresponding depth image can be restored at the same time, and the method conforms to the actual application scene.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a dual-stream network-based RGBD image joint restoration method. Background technique [0002] In the era of informatization and big data, image information is used in all aspects of life. People can easily obtain digital images around them through mobile phones, cameras and other sensor devices, so as to spread and share them. In recent years, with the development of visual sensors and advanced technologies such as artificial intelligence, people's requirements for describing and disseminating visual information around them are no longer satisfied with general 2D information. Rich and colorful 3D visual information with better expressive ability has gradually entered people's minds. life. Usually 3D visual information is obtained by RGBD data sensors such as Microsoft Kinect and further processed and modeled. However, the processing capability of RGBD data sens...

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): G06F16/51G06T5/00G06N3/04
CPCG06F16/51G06T5/007G06T2207/20081G06T2207/10024G06N3/045
Inventor 许勇祝叶李芃
Owner SOUTH CHINA 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
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