Unlock instant, AI-driven research and patent intelligence for your innovation.

Multi-scale point cloud classification implementation method based on graph convolution

An implementation method, multi-scale technology, applied in instruments, biological neural network models, character and pattern recognition, etc., can solve problems such as lack of multi-scale features

Active Publication Date: 2021-04-09
HUBEI UNIV OF TECH
View PDF4 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method can only capture edge features in a limited neighborhood and lacks multi-scale features
Therefore, it is still challenging to further improve the performance of 3D point cloud classification

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
  • Multi-scale point cloud classification implementation method based on graph convolution
  • Multi-scale point cloud classification implementation method based on graph convolution
  • Multi-scale point cloud classification implementation method based on graph convolution

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056]Referfigure 1 ,figure 2 withimage 3The method of implementing a multi-scale point cloud classification of the present invention based on a map of the present invention includes:

[0057]Step 1, improve the KNN neighboring algorithm, get the KN points of the equal interval sampling within a different scale range, thereby constructing a m-knn diagram of different scales of the point cloud collection:

[0058]A D-Diming Cloud with N points is expressed as:

[0059]

[0060]Where x represents a point cloud collection, XiIndicates that each point, n represents the number of points in the point cloud collection, and D represents the dimension of the point cloud data.

[0061]Since only the location information of the point cloud is used in the present invention, ie D = 3. Therefore, each point contains only its three-dimensional coordinate, namely:

[0062]Xi= (XiYi,zi) (2)

[0063]According to the coordinate calculation point cloud collection, the European distance between:

[0064]

[0065]Calculate points ...

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 belongs to the technical field of three-dimensional point cloud classification, and discloses a multi-scale three-dimensional point cloud classification implementation method based on graph convolution. The method aims to obtain richer neighborhood information by extracting edge features of different scales of a point cloud set so as to improve the accuracy of point cloud classification. The method comprises the following specific steps: step 1, acquiring k points sampled at equal intervals in different scale ranges through an improved KNN proximity algorithm so as to construct M-KNN images of different scales of a point cloud set; and step 2, constructing an EdgeConv (edge convolution layer) module by using the M-KNN graph to realize extraction of point cloud edge features. and step 3, constructing a multi-scale point cloud classification network model by using the edge convolution layer, and training the network by using the ModelNet40 data set to obtain a final point cloud classification network.

Description

Technical field[0001]The present invention belongs to the field of three-dimensional dot cloud classification, and more particularly to a multi-scale three-dimensional point cloud classification implementation method based on the map volume.Background technique[0002]The three-dimensional dot cloud classification refers to the process of extracting artificial or natural geographic elements from complex callback clouds with three-dimensional spatial coordinates. In recent years, with the development of three-dimensional laser scanning technology, the adoption of three-dimensional dot cloud data has also become rapid and cheap. There is a very wide application in the fields of unmanned, robots, indoor scene detection and identification, and the three-dimensional point cloud classification has become a research hotspot in the computer visual field.[0003]The traditional point cloud classification method is to classify the three-dimensional dot clouds by manually designing a series of fea...

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): G06K9/62G06N3/04
CPCG06N3/045G06F18/214G06F18/24147
Inventor 张正文胡永东巩朋成张林让李婕潘懋舜柯凡余梦婕
Owner HUBEI UNIV OF TECH