Processing Feature Recognition Based on Improved YOLOv5 Algorithm

Author:

Miao Liqin,Li Jialong1,Feng Xin,Tang Zhengwei,Liang Huaidan

Affiliation:

1. Changchun University of Science and Technology

Abstract

Abstract

Processing feature recognition technology can effectively integrate CAD and CAPP systems, thus realizing the effective interoperability of design information and manufacturing information. Classical processing feature recognition technology does not have the ability to learn, recognition efficiency is low and the accuracy of the recognition results can not be guaranteed, while the deep neural network technology not only has the ability to learn, and noise resistance can improve the accuracy of the recognition, therefore, the proposed deep learning-based processing feature classification and recognition technology, the use of multi-angle downscaling to capture the processing features of the processing features of the image method to create a processing feature dataset, and the deep neural network model through the YOLOv5 improved deep neural network model to train the dataset, and use the machining feature classifier to classify the processed part model and output the machining feature results. The machining feature recognition process is verified with typical parts as an example by combining with UG platform, which provides a new idea for machining feature recognition technology.

Publisher

Research Square Platform LLC

Reference28 articles.

1. Ruhuan X (2019) Hybrid algorithm for processing feature recognition based on geometric inference and neural network [D]. Huazhong University of Science and Technology

2. Graph-based heuristics for recognition of machined features from a 3D solid model[J];Joshi S;Comput Aided Des,1988

3. High-level feature recognition using feature relationship graphs[J];Huang Z;Comput Aided Des,2002

4. Freeform Machining Feature Recognition with Manufacturability Analysis[J];Cai N;Procedia CIRP,2018

5. Woo TC (1982) Feature Extraction by Volume Decomposition[C]//In Pro CAD/CAM Technology Report in Mechanical Engineering MIT

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3