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

Three-dimensional image processing method

A technology of three-dimensional image and processing method, which is applied in the field of image processing, and can solve the problems of calculation waste of 0 element operation, good convolution effect, and inability to obtain

Inactive Publication Date: 2018-05-29
FAFA AUTOMOBILE (CHINA) CO LTD
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] a. Since the laser point cloud data is sparse, most of the calculations in this calculation process are wasted on the operation of 0 elements;
[0007] b. The convolutional network of three-dimensional data needs to determine the adjacent units of each element, but the determination of the neighbors in three-dimensional space is very time-consuming compared with images;
[0008] c. For the convolution operation of each element, it is also necessary to know its geometric features, color features and other information. Only the geometric position information of the point cloud cannot obtain a better convolution effect.

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
  • Three-dimensional image processing method
  • Three-dimensional image processing method
  • Three-dimensional image processing method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0099] like figure 1 Shown is a working flow chart of a three-dimensional image processing method provided by an embodiment of the present invention, including:

[0100] Step S101, converting the three-dimensional image into a corresponding octree structure;

[0101] Step S102, performing a convolution operation on the octree structure to obtain an octree structure after the convolution operation.

[0102] An octree is a tree-like data structure used to describe a three-dimensional space. like image 3 As shown, each node 31 of the octree represents a volume element of a cube, and each node has eight child nodes, and the volume elements represented by the eight child nodes are equal to the volume of the parent node. Generally, the center point is used as the bifurcation center of the node. Since the octree is only composed according to the existing point cloud information, it can avoid a lot of 0-element space; on the other hand, the octree is expressed in a tree structure...

Embodiment 2

[0105] like figure 2 The embodiment of the present invention provided as an optional embodiment of the present invention provides a workflow flowchart of a three-dimensional image processing method, including:

[0106] Step S201, according to the positional relationship of each pixel in the 3D image, determine the tree structure of the octree structure, and determine the feature information vector of each leaf node.

[0107] Specifically,

[0108] The feature information vector is calculated based on the K nearest neighbor point cloud of each leaf node corresponding to the pixel, and K is a natural number greater than 1.

[0109] The feature information vector can include feature root information (line, surface, and volume feature root information), curvature information, and reflection rate information indicating the surface characteristics of the object, so that each leaf node is converted into a 5-dimensional vector. Furthermore, the initial input data is converted into ...

Embodiment 3

[0131] like Figure 4 A device block diagram of a three-dimensional image processing device provided by an embodiment of the present invention, including:

[0132] A conversion module 401, configured to: convert a three-dimensional image into a corresponding octree structure;

[0133] The convolution module 402 is configured to: perform a convolution operation on the octree structure to obtain an octree structure after the convolution operation.

[0134] In the embodiments of the present invention, a large amount of space can be saved by converting a three-dimensional image into an octree structure, and at the same time, a large amount of computing time can be saved, efficiency can be improved, and convolution operations on three-dimensional graphics can be realized.

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 embodiment of the invention discloses a three-dimensional image processing method. The method comprises the steps that a three-dimensional image is converted into the corresponding octree structure; and convolution operation is performed on the octree structure so that the octree structure after convolution operation can be obtained. The three-dimensional image is converted into the octree structure so that a lot of space can be saved, a lot of operation time can be saved, the efficiency can be enhanced and convolution operation of the three-dimensional image can be realized.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a three-dimensional image processing method. Background technique [0002] Environmental perception is the key technology that determines the ability of unmanned vehicle systems to recognize environmental information. Laser (Velodyne 64e)-based environmental perception algorithms can effectively extract road drivable areas, object recognition and tracking, etc., but traditional laser-based point cloud The scene perception algorithm of the data needs to simply classify the scene by designing a supervised classification and segmentation method based on the geometric feature information of the 3D point cloud data. [0003] However, the inventor found in the process of realizing the invention that the existing convolution calculation method cannot be applied to three-dimensional image processing. [0004] The convolutional network of the traditional two-dimensional image us...

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): G06K9/00G06K9/62
CPCG06V20/56G06F18/24133
Inventor 殷鹏
Owner FAFA AUTOMOBILE (CHINA) CO LTD
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