A block patch reconstruction method for point cloud data based on deep learning

A technology of point cloud data and deep learning, which is applied in the field of block face reconstruction of point cloud data, can solve problems such as result errors, achieve the effects of avoiding geometric operations, powerful fitting capabilities, and improving reliability

Active Publication Date: 2021-10-29
ZHEJIANG UNIV
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For example, the classic Poisson surface reconstruction method is very dependent on the normal information of the point cloud. If the normal information is noisy, it will cause large errors in the generated results.
The recently proposed DeepSDF (J.J.Park et al.ICCV2019) method is also a 3D point cloud reconstruction method based on deep learning. This method does not rely on the normal information of the point cloud, but this method can only be used for individual categories of models, such as airplanes and sofas. etc. for 3D reconstruction

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 block patch reconstruction method for point cloud data based on deep learning
  • A block patch reconstruction method for point cloud data based on deep learning
  • A block patch reconstruction method for point cloud data based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0025] The present invention will be described in detail below according to the accompanying drawings.

[0026] refer to figure 1 , the deep convolutional neural network used in the present invention needs to use the voxel information converted from the point cloud as input, and output the corresponding SDF. Therefore, during training, this method prepares the matching voxel information and SDF as the data set for network training. This method uses the original patch data in the A Benchmark for 3D Mesh Segmentation dataset open sourced by Princeton University to calculate the corresponding point cloud data containing normal information, and generate the corresponding SDF. Randomly select a point in the generated point cloud as the center of the circle, search for adjacent points in the fixed-size radius and generate voxels as the input data of the deep convolutional neural network, and then select SDF in the corresponding area as the true value for training .

[0027] figur...

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 presented in this paper is called a block-based surface reconstruction method from point cloud data based on deep learning. The invention discloses a method for reconstructing a surface model from a point cloud based on deep learning. The method utilizes point cloud data in three-dimensional space to generate SDF (Signed Distance Field, directed distance field) in blocks and integrates each block to obtain a complete SDF, and finally use the Marching Cubes algorithm to get the final patch data. The present invention can still perform robustly in the presence of noise in the point cloud data, especially in the case of normal information deviation, greatly reducing the requirement for the direction accuracy of the collected point cloud data; during operation, the present invention can also process in parallel ,efficient. As for the application of the present invention, it mainly focuses on the field of three-dimensional object reconstruction, and has wide application space in three-dimensional modeling in digital entertainment, computer-aided design and the like.

Description

technical field [0001] The invention belongs to the fields of computer graphics and artificial intelligence, and in particular relates to a block face reconstruction method based on deep learning point cloud data. Background technique [0002] 3D reconstruction has been widely used in the fields of digital entertainment and computer-aided design in recent years; virtual reality technology, enhanced display technology, 3D animation movies, map imaging, etc. all require a large number of 3D models. If all these 3D models need to be designed manually, it will consume a huge amount of human resources, but the current fully automatic 3D reconstruction technologies based on pictures, point clouds, etc. have their own defects. [0003] The present invention is aimed at the three-dimensional patch reconstruction work using point cloud data. There are some defects in the previous 3D reconstruction work of point cloud. For example, the classic Poisson surface reconstruction method r...

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): G06T17/00G06N3/04G06N3/08
CPCG06T17/00G06N3/08G06N3/045
Inventor 郑友怡
Owner ZHEJIANG UNIV
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