An end-to-end method for high-precision shape modeling of industrial parts

A modeling method and high-precision technology, applied in neural learning methods, biological neural network models, image analysis, etc., can solve the problem that the shape modeling of industrial parts cannot adapt to various data environments, and shorten algorithm debugging. Time, strong adaptability and accuracy, simple model debugging effect

Active Publication Date: 2020-10-30
视研智能科技(广州)有限公司
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The present invention provides an end-to-end high-precision shape modeling method for industrial parts in order to overcome the defect that the shape modeling of industrial parts described in the prior art cannot adapt to various data environments

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
  • An end-to-end method for high-precision shape modeling of industrial parts
  • An end-to-end method for high-precision shape modeling of industrial parts

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] This embodiment provides an end-to-end high-precision shape modeling method for industrial parts. Such as figure 1 As shown, it mainly includes: S1: edge extraction, S2: building topological relationship between points, S3: point feature extraction, S4: point optimization. The four modules will be described separately below.

[0045] S1: Initial edge point extraction.

[0046] S1.1: Generate a depth map using the input point cloud.

[0047] S1.2: Mark the area of ​​each part and its internal edge structure on the depth map as the ground truth for model training.

[0048] S1.3: Use Mask-RCNN on the depth map for target-level detection and semantic segmentation, and at the same time predict the distance between each pixel and its nearest edge point, referred to as the distance map.

[0049] S1.4: Obtain the area of ​​each target detected by S1.3. In the distance map for this region, set pixel values ​​less than 2 to 1, and to 0 otherwise. The result is the initial e...

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 relates to an end-to-end high-precision shape modeling method for industrial parts, including: S1: obtaining an edge map through edge point extraction; S2: constructing a topological relationship between points according to the edge map; S3: performing a point map belonging to each target Extract feature points; S4: Encode-decode point optimization for feature points to achieve model parameter optimization. The method of the present invention has strong adaptability and accuracy, and can stably extract concise and accurate shape lines under the conditions of data noise, occlusion and shadow; has great adaptability, and can be directly applied to various types of parts geometric modeling; the method of the present invention effectively solves the problem that the shape and lines cannot be stably extracted under the conditions of data noise, occlusion and shadow; at the same time, the model debugging is simple, and compared with the traditional heuristic shape modeling technology, the time is greatly shortened. Algorithm debugging time.

Description

technical field [0001] The present invention relates to the technical field of industrial part modeling, and more specifically, relates to an end-to-end high-precision industrial part shape modeling method. Background technique [0002] The shape modeling technology of industrial parts is one of the core technologies of robot welding, detection and measurement of industrial parts. Existing methods generally use various line feature or point feature extraction operators, such as Hough line feature operator and Sift, Harris, local gradient and other point feature extraction operators. Then design heuristic rules based on the shape characteristics of the target, connect adjacent broken lines, and remove errors such as misdetected line segments to achieve shape modeling. Point and line features can also be used directly to guide robotic welding and template matching with CAD models. These operators are often unable to adapt to various data environments, and are prone to false ...

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): G06T7/00G06T7/13G06T7/136G06T7/187G06T5/30G06N3/04G06N3/08
CPCG06T7/0004G06T7/13G06T7/136G06T7/187G06T5/30G06N3/084G06T2207/10028G06T2207/20081G06T2207/20084G06T2207/30164G06N3/045
Inventor 王磊吴伟龙周建品李争
Owner 视研智能科技(广州)有限公司
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