Supercharge Your Innovation With Domain-Expert AI Agents!

Robot depth camera point cloud downsampling filtering method based on slicing and interpolation

A depth camera and robot technology, applied in the direction of instruments, image acquisition, processor architecture/configuration, etc., can solve the problems of point cloud missing feature points, complex algorithm implementation, large computing resources, etc., to reduce CPU usage and improve system performance , the effect of small data storage

Inactive Publication Date: 2021-02-02
JINAN INSPUR HIGH TECH TECH DEV CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Both methods will cause the point cloud to lose feature points. At the same time, down-sampling methods such as voxel filtering will mechanically filter out some points, resulting in point cloud distortion. In addition, these methods need to traverse all point clouds, and the algorithm implementation is more complicated and consumes massive computing resources

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
  • Robot depth camera point cloud downsampling filtering method based on slicing and interpolation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032]In step a), the point cloud data collected by the ARM platform at the sending end of the robot is input to the FPGA processor, and the Verilog hardware description language is used in the FPGA processor to perform digital logic involving various functional modules to achieve point cloud downsampling filtering, data transmission, and receiving endpoints Cloud interpolation restore function.

Embodiment 2

[0034]In step a), the FPGA processor at the receiving end of the robot restores the point cloud and sends it to the FPGA processor at the receiving end.

Embodiment 3

[0036]When the FPGA processor performs the slicing of the point cloud data, it uniformly randomly samples the N point cloud machine radius in the slice in the vicinity of the point within r, and finds the number of consecutive point clouds in the neighborhood that exceeds k and the outer contour curvature is greater than q Neighborhood, calculate the number of sub-neighborhoods c, count the total number f of slices where c is less than s, when the total number f is less than t and consecutive e frames appear at the same time, increase the slice span by 1, when the total number When f is greater than u and consecutive e frames appear simultaneously, the slice span is reduced by 1. N is the total number of point clouds in the slice / 1000, r=0.01m, k=5, outer contour curvature=distance of adjacent outer contour point clouds / 0.01, q=0.11, c is a positive integer, f:s=N / 10, f is a positive integer, t=4, e=3, u=6.

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 a robot depth camera point cloud downsampling filtering method based on slicing and interpolation. The method is matched with straight-through filtering, point cloud data obtained after straight-through filtering are circularly sliced according to the x axis, the y axis and the z axis in the time sequence, planes are selected alternately in a spacer mode, odd pieces are scanned for the previous time, the odd pieces form an odd field, even pieces are scanned for the next time, and even pieces form an even field. Alternate scanning, slice taking and publishing or transmission are carried out, and only single-field data are transmitted in the point cloud transmission process. And when receiving the point cloud data and processing or applying the point cloud data, thereceiving terminal interpolates the single-field data to restore the whole frame of point cloud data. The algorithm is easy to implement, small in calculation amount and suitable for robot embedded platform processing. The method is low in data storage capacity, capable of performing pipeline operation and high in real-time performance, only needs to cache one piece of point cloud in a single time, and is suitable for parallel operation platforms such as an FPGA or a GPU.

Description

Technical field[0001]The invention relates to the field of robots and embedded systems, in particular to a point cloud down-sampling filtering method of a robot depth camera based on slicing and interpolation.Background technique[0002]Traditional point cloud filtering methods mainly include straight-through filtering, voxel filtering, etc., straight-through filtering is to crop according to scale coordinates, select the required range of point clouds, one is point cloud downsampling, and directly filter all point clouds to reduce point cloud density degree. Both methods will cause the point cloud to lose characteristic points, and down-sampling methods such as voxel filtering will mechanically filter out some points, causing point cloud distortion. In addition, this type of method needs to traverse all point clouds, and the algorithm implementation is more complicated and costly. A lot of computing resources.Summary of the invention[0003]In order to overcome the shortcomings of the ...

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): G06T1/20G06T1/60G06T1/00
CPCG06T1/20G06T1/60G06T1/0007
Inventor 崔欣高明马辰董崇良
Owner JINAN INSPUR HIGH TECH TECH DEV CO LTD
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com