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

A high-precision disparity calculation acceleration method based on GPU

A parallax calculation and high-precision technology, applied in the field of 3D reconstruction, can solve problems such as large amount of calculation, achieve the effect of improving effectiveness, improving effectiveness and parallelism, and reducing area size

Active Publication Date: 2020-09-15
洛阳中科信息产业研究院 +1
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, to obtain high-precision disparity usually means to bear a large amount of calculation

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 high-precision disparity calculation acceleration method based on GPU
  • A high-precision disparity calculation acceleration method based on GPU
  • A high-precision disparity calculation acceleration method based on GPU

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] A GPU-based method for accelerating high-precision parallax calculations. The GPU is set to a unified computing device architecture. The specific steps include.

[0033] S1. Obtain left and right perspective images. Both left and right perspective images are infrared images, and each pixel on the left and right perspective images is randomly assigned a current disparity value and based on the disparity value Calculate the current disparity number.

[0034] S2. Split the left-view image and the right-view image into several 32×32 propagation blocks, and there is an overlap area between any two adjacent propagation blocks, and the overlap area is arranged in 32 Rows of pixels or 32 pixels arranged in columns.

[0035] S3. Iteratively update the disparity value of each propagation block in the left-view image and the right-view image, that is, correct the disparity value of each pixel in the left-view image and the right-view image by iterative updating. In the iterative update...

Embodiment 2

[0047] In order to further speed up the disparity calculation, the present invention provides a second embodiment. On the basis of the first embodiment, the specific steps include: S1', obtaining a left view image, a right view image and a reference image; S2', extracting the target area from the reference image, shifting the target area to the right to obtain the left reference image, The target area is shifted to the left to obtain the right reference image; S3', the part of the left view image corresponding to the target area on the left reference image is set as the left effective area, and the right view image is compared with the target area on the right reference image. The corresponding part is set as the right effective area; S4', the left effective area and the right effective area are divided into several effective blocks, the size of the effective block is an integral multiple of 32×32; S5', the effective block is split into Several 32×32 propagation blocks, iterati...

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 high-precision parallax computation acceleration method based on a GPU. The GPU is set as unified computing equipment architecture, and the method comprises the specific stepas follows: S1, acquiring a left viewing angle image and a right viewing angle image; S2, respectively splitting the left viewing angle image and the right viewing angle image into a plurality of propagation blocks in size of 32*32; S3, performing parallax value iteration updating of each propagation block in the let viewing angle image and the right viewing angle image, saving all pixel values of the propagation block in the parallax value iteration updating in a texture memory of the GPU, and saving a process parameter in the parallax value iteration updating process in a shared memory of the GPU. Through the method disclosed by the invention, the computing capacity of the CPU can be sufficiently utilized, and the computing speed is greatly accelerated in the premise of guaranteeing theparallax computing precision.

Description

Technical field [0001] The invention relates to the technical field of three-dimensional reconstruction, in particular to a GPU-based high-precision disparity calculation acceleration method. Background technique [0002] As an important branch of computer vision, the acquisition process of binocular stereo vision is roughly by a camera that shoots the same scene by moving or rotating the angle of view, or two cameras that simulate binoculars to shoot the same object, and then use a stereo matching algorithm Calculate the parallax between the corresponding image points of the two images, combine the parallax image and the camera calibration parameters to obtain the three-dimensional coordinates of each point of the measured object in the scene, and then reconstruct the three-dimensional structure information of the scene and obtain the depth value of the corresponding point. The depth value is the actual distance between the imaging plane and the measured object. A complete ster...

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): H04N13/128H04N13/239
Inventor 冯文铎
Owner 洛阳中科信息产业研究院
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