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

Semantic segmentation method of 3D point cloud data based on deep learning and self-attention

A technology for point cloud data and semantic segmentation, applied in image data processing, image analysis, character and pattern recognition, etc., can solve the problem of low segmentation accuracy and achieve the effect of improving accuracy

Active Publication Date: 2021-09-03
XIDIAN UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the defects in the above-mentioned prior art, and propose a 3D point cloud data semantic segmentation method based on deep learning and self-attention mechanism, which is used to solve the low segmentation precision existing in the prior art question

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
  • Semantic segmentation method of 3D point cloud data based on deep learning and self-attention
  • Semantic segmentation method of 3D point cloud data based on deep learning and self-attention
  • Semantic segmentation method of 3D point cloud data based on deep learning and self-attention

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] Refer figure 1 The present invention includes the following steps:

[0033] Step 1) Get the training set R 2 And verification set V 2 :

[0034] Step 1A) Get the 3D point cloud data file with the label from the database f: {f 1 , F 2 , ..., f i , ..., f f ,}, And proportion there is N R 3D point cloud data file as an initial training set R 0 , The rest f (1-N R 3D point cloud data file as an initial verification set V 0 , F i Indicates the i-th 3D point cloud data file, f is the total number of 3D point cloud data file, f ≥ 100, 0.6 ≤ N R R = 0.8, this will randomly select 80% of 3D point cloud data files from the database as the initial training set R 0 , The remaining 20% ​​3D point cloud data files as the initial verification set V 0 ;

[0035] Step 1B) Put R 0 Enter the PDAL library for blocking to get the training data block set R...

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 present invention proposes a 3D point cloud data semantic segmentation method based on deep learning and self-attention mechanism, which is used to solve the technical problem of low segmentation accuracy existing in the prior art. The implementation steps include: (1) Obtaining a training set and verification set; (2) construct the 3D point cloud data semantic segmentation network of deep learning and self-attention mechanism; (3) set the loss function required for the 3D point cloud data semantic segmentation network of training deep learning and self-attention mechanism; (4) Perform supervised training on the 3D point cloud data semantic segmentation network with deep learning and self-attention mechanism; (5) Obtain the semantic segmentation results of the 3D point cloud data test set. The present invention adds a self-attention module to the deep learning network, which can better extract deep features including the relationship between various feature channels, thereby improving the segmentation accuracy.

Description

Technical field [0001] The present invention belongs to the field of radar 3D dot cloud data processing, and it is directed to a 3D point cloud data segmentation method, which specifically relates to a semantic semantic semantic semantic semantic division method based on deep learning and self-focus mechanism. Can be used in automatic driving, robotics, 3D maps, land surveying, prospect segmentation, smart city construction, agricultural estimates, forestry resource census, ecological environment monitoring and disaster prevention and mitigation. Background technique [0002] In recent years, with the development of deep sensors, point cloud processing has become one of the research hotspots. Point cloud data refers to: Scanning information is recorded in the form of a point, and each point contains three-dimensional coordinates, and some may contain information such as color information, reflection intensity information, gray value, depth, or the number of returns. Generally, it...

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 Patents(China)
IPC IPC(8): G06K9/62G06T7/10
CPCG06T7/10G06T2207/10028G06F18/211G06F18/214
Inventor 焦李成李玲玲张杰张格格马清华郭雨薇丁静怡张梦璇程曦娜王佳宁
Owner XIDIAN UNIV
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