Point cloud data registration method, system and device and computer storage medium

A technology of computer storage and point cloud data, applied in computing, image data processing, neural learning methods, etc., can solve problems such as insufficient robustness, and achieve good robustness

Active Publication Date: 2020-08-11
SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
View PDF7 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Among them, the global registration method is to obtain the transformation relationship between the two frames of point clouds and their relative positional relationship when the initial state of the two frames of point clouds is unknown. Obviously, this type of method is more versatile, but the current global Registration algorithms are mostly based on the local features of the point cloud, using the descriptor distance as the metric to match two frames of point clouds. Such methods rely heavily on feature extraction and matching algorithms, that is, rely on the low-level visual features of the point cloud , easily affected by environmental factors such as occlusion and other factors such as point cloud density, not robust enough

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
  • Point cloud data registration method, system and device and computer storage medium
  • Point cloud data registration method, system and device and computer storage medium
  • Point cloud data registration method, system and device and computer storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] refer to figure 1 , the registration method of the point cloud data in the present embodiment comprises steps:

[0056] S1. Obtain a first point cloud dataset and a second point cloud dataset.

[0057] Specifically, the first point cloud data set and the second point cloud data set respectively correspond to two images under different viewing angles in the target space, and the two images partially overlap, that is, the first point cloud data set and the second point cloud data set have overlapping area. The data points in the first point cloud dataset and the second point cloud dataset are all three-dimensional data points, and the first point cloud dataset and the second point cloud dataset can be directly obtained by a depth image sensor, for example, directly scanned by a laser radar can also be obtained after conversion through the depth image conversion method.

[0058] S2. Extract semantic information of the first point cloud dataset and the second point cloud...

Embodiment 2

[0104] The difference between this embodiment and Embodiment 1 is that the first point cloud data set and the second point cloud data set in this embodiment include multiple objects respectively, and the categories of multiple objects are the same, because these objects cannot be classified by category Therefore, the semantic information in this embodiment is the shape of the object. The object in this embodiment can have a grid-like or other similar grid-like repeating structure, or it can be an object with a specific shape, such as , spheres, cuboids, cylinders, cones, etc., when extracting semantic information, you can obtain repetitive structures of the same category or objects with specific shapes by extracting based on repetitive structures or special structures.

[0105] In this embodiment, except that steps S2 and S4 are different from those in Embodiment 1, other steps are the same as in Embodiment 1 and will not be repeated here. Steps S2 and S4 in this embodiment wil...

Embodiment 3

[0120] refer to Figure 7 , this embodiment also provides a point cloud data registration system, the registration system includes a first acquisition module 1 , an extraction module 2 , a second acquisition module 3 , a registration module 4 , and a conversion module 5 .

[0121] Specifically, the first acquisition module 1 is used to acquire the first point cloud data set and the second point cloud data set, and the extraction module 2 is used to respectively extract the semantic information of the first point cloud data set and the second point cloud data set. The second acquisition module 3 is used to obtain the semantic connection graphs of the first point cloud data set and the second point cloud data set respectively according to the semantic information of the first point cloud data set and the second point cloud data set, and the registration module 4 is used for The semantic connection graph of the first point cloud data set and the second point cloud data set is reg...

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 provides a point cloud data registration method, system and device and a computer storage medium. The registration method comprises the steps: obtaining a first point cloud data set anda second point cloud data set; respectively extracting semantic information of the first point cloud data set and the second point cloud data set; respectively obtaining semantic connection graphs ofthe first point cloud data set and the second point cloud data set according to the semantic information of the first point cloud data set and the second point cloud data set; registering the semanticconnection graphs of the first point cloud data set and the second point cloud data set to obtain a first transformation relationship between the first point cloud data set and the second point clouddata set; and converting the second point cloud data set according to the first conversion relationship. The point cloud data registration method provided by the invention starts from high-level semantic information, is not limited by low-level visual characteristics of the point cloud, is not easily influenced by other factors such as environmental factors and point cloud density, and has good robustness.

Description

technical field [0001] The present invention relates to the technical field of point cloud data registration, in particular to a point cloud data registration method, system, device and computer storage medium. Background technique [0002] With the rapid development of information technology and the increasing demand for 3D models in various industries such as movies and games, the acquisition of 3D models has become more and more important. In the field of computer vision and computer graphics, 3D reconstruction refers to the process of obtaining a 3D model from the captured shape and appearance of real objects. The representation of the 3D model includes point cloud, voxel, grid, etc. The point cloud is a collection of 3D points, in other words, it uses a collection of 3D points to represent the 3D model. Point cloud is a very important data form. It can directly obtain 3D point cloud through laser radar and other equipment; it can also use depth camera to obtain depth i...

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): G06T7/30G06N3/04G06N3/08
CPCG06T7/30G06N3/08G06T2207/10012G06N3/045
Inventor 程章林周佳新奥利夫·马丁·多伊森
Owner SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
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