A point cloud registration method based on extended Gaussian image

An extended image and point cloud registration technology, applied in image enhancement, image analysis, image data processing, etc., can solve the problems of long time consumption, difficulty in balancing efficiency and accuracy, and high cost of SAC-IA algorithm, and achieve time-consuming Short, improve division efficiency, avoid local optimization effect

Active Publication Date: 2019-02-15
HARBIN INST OF TECH
View PDF3 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the excessive cost of the coarse registration algorithm SAC-IA algorithm commonly used in the field of existing point cloud registration, the Gaussian sphere is divided into polyhedrons or latitude and longitude based on the EGI algorithm, which takes a long time, and the calculation based on the EGI algorithm When rotating the matrix, it is difficult to directly search for the optimal solution in the feasible region, and it is difficult to balance the dual requirements of efficiency and accuracy. A point cloud registration method based on extended Gaussian images is proposed

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 point cloud registration method based on extended Gaussian image
  • A point cloud registration method based on extended Gaussian image
  • A point cloud registration method based on extended Gaussian image

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0030] Specific implementation mode one: combine figure 1 Describe this embodiment, a kind of point cloud registration method based on extended Gaussian image of this embodiment specific process is:

[0031] Step 1. Record the coordinate sets of the normals in the model point cloud and blank point cloud in their respective body coordinate systems as N T and N S (Get the normal position);

[0032] The model point cloud is a point cloud of an ideal mechanical part model;

[0033] The blank point cloud is the point cloud of the blank (the blank processed by the factory and scanned by the scanner);

[0034] Step 2, respectively meshing the extended Gaussian image (EGI) generated by the model point cloud and the blank point cloud;

[0035] Step 3. Based on step 2, pair grid P i T , and grid P i s , Perform preliminary matching, and the grid P after preliminary matching i T , and grid P i s , Sort in descending order according to the number of normals in the grid...

specific Embodiment approach 2

[0047] Specific embodiment two: the difference between this embodiment and specific embodiment one is that in said step two, the Gaussian extended image (EGI) generated by the model point cloud and the blank point cloud is respectively meshed and divided, and the specific process is:

[0048] HEALPix (Hierarchical Equal Area isoLatitude Pixelization) was originally applied to the research of the cosmic microwave background to support the discretization of spherical high resolution. HEALPix can be divided into several levels according to the resolution, the lowest level is 12 grids. HEALPix divides the sphere as figure 2 , 3 , 4, and 5.

[0049] HEALPix has the following characteristics:

[0050] The grid blocks are structured hierarchically and the data is easily accessible.

[0051] The method of generating grid blocks is simple and the operation is efficient.

[0052] The sphere is divided into equal-sized surface quadrilaterals, each meshing with the same area.

[00...

specific Embodiment approach 3

[0059] Specific embodiment 3: The difference between this embodiment and specific embodiments 1 or 2 is that in the step 3, the pairing of the grid P based on the step 2 i T , and grid P i s , Perform preliminary matching, and the grid P after preliminary matching i T , and grid P i s , Sort in descending order according to the number of normals in the grid, denoted as P 1 T″ , and P 1 s″ , The specific process is:

[0060] Correspondence point identification

[0061] Finding the corresponding mesh can be a challenging job, especially when the object's normals are somewhat evenly distributed. Due to the distribution of normal vectors, a grid with a high density may have a grid with a similar density in the opposite direction, causing errors in the calculation of the rotation matrix. Also, a good algorithm should be insensitive to rotation and position changes of the mesh. Here we use a simple yet effective method to find a mesh in the target EGI that i...

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

A point cloud registration method based on an extended Gaussian image relates to a point cloud registration method of an image. The purpose of the invention is to solve the problem of the coarse registration algorithm SAC commonly used in the prior point cloud registration field. The overhead of IA algorithm is too large. It takes a long time to divide the Gaussian sphere into polyhedron or longitude and latitude based on EGI algorithm, and it is difficult to balance the efficiency and precision requirements when searching the optimal solution in feasible domain directly based on EGI algorithm. The procedure is as follows: the coordinate set of the normal in the respective ontology coordinate system is recorded as NT and NS; Gaussian extended image is gridded. Sorting the preliminarily matched grids in descending order according to the number of normal lines in the grids; The modified mesh is obtained; Calculating a rotation matrix; Output iter, Corr and M. The present invention is used in the intersecting field of computer graphics and machining technology.

Description

technical field [0001] The invention belongs to the intersecting field of computer graphics and mechanical processing technology, and specifically relates to a point cloud registration method of images. Background technique [0002] In the process of workpiece processing, casting is often used to manufacture rough blanks first, and then the parts are processed into final shapes by subtractive processing methods such as turning and grinding. In this process, it is necessary to detect the condition of the casting, determine the distribution of the allowance, and then further optimize the distribution of the allowance to facilitate the next cutting process. Laser scanners are more suitable for the inspection of castings than manual labor because of their rapid surface point cloud data, low time cost, and high precision. After the scanner obtains the point cloud data of the component, the point cloud registration algorithm is used to register and compare the point cloud with th...

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): G06T7/30
CPCG06T7/30G06T2207/10028
Inventor 杜志江高永卓董为徐威李明洋
Owner HARBIN INST OF TECH
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