Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Depth image super-resolution method based on multidirectional dictionary

A depth image and super-resolution technology, applied in image analysis, image enhancement, image data processing, etc., can solve the problems of low quality of results, blurred texture boundary area, etc., to achieve resolution improvement, good spatial smoothness, and texture boundary. Clear details

Active Publication Date: 2018-03-06
BEIJING UNIV OF TECH
View PDF10 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] To sum up, the existing depth image super-resolution algorithm obtains low-quality results, and the texture boundary area is relatively blurred, which has certain limitations.

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
  • Depth image super-resolution method based on multidirectional dictionary
  • Depth image super-resolution method based on multidirectional dictionary
  • Depth image super-resolution method based on multidirectional dictionary

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0069] In order to more clearly describe the technical contents of the present invention, further describe below in conjunction with specific examples:

[0070] The frame diagram of the present invention is as figure 1 , the specific implementation process is divided into two stages, the dictionary learning stage and the super-resolution reconstruction stage.

[0071] 1. Dictionary learning stage

[0072] The dictionary learning phase is divided into four steps: collecting training image sets, image feature extraction, computing geometric orientations, and dictionary training.

[0073] 1. Collect training images

[0074] First, collect enough high-resolution depth images and corresponding color images from standard images as the training set for dictionary training, such as figure 2 , where the low-resolution depth image is down-sampled from the high-resolution depth image.

[0075] 2. Image feature extraction

[0076] a. Color image contour extraction:

[0077] Interpo...

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 depth image super-resolution method based on a multidirectional dictionary. The method comprises the following steps: 1) extracting color image features and obtaining a profile diagram; 2) carrying out interpolation on a low-resolution depth image, and expressing features thereof through high-pass filtering; 3) carrying out block segmenting on the low-resolution depth image, and calculating geometric direction of each image block; 4) carrying out dictionary training; and 5) selecting a dictionary according to the geometric direction obtained in the step 3) and reconstructing a depth image. According to the technical scheme, resolution of the depth image is allowed to be enhanced, and boundary texture region is clear.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to a method for super-resolution of depth images based on a multi-directional dictionary. Background technique [0002] In recent years, with the rapid development of time-of-flight cameras and Microsoft Kinect and other 3D cameras, depth images have been widely used in low-cost computer vision fields, such as robot navigation, augmented reality, and scene analysis. However, limited by the external conditions and the equipment itself, the depth images obtained by using these 3D cameras often have problems of low resolution, a lot of noise, and the loss of some depth boundary structures, so these depth images cannot be directly used for depth perception and 3D reconstruction. [0003] Depth image super-resolution algorithms have made great progress, which are mainly divided into two aspects: optimization-based methods and filter-based methods. Optimization-based methods can...

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): G06T5/50G06T3/40G06T7/90
CPCG06T3/4053G06T5/50G06T7/90G06T2207/10024G06T2207/20024
Inventor 王瑾许伟王志强朱青
Owner BEIJING UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com