首页|基于型材面结构特征匹配方法研究

基于型材面结构特征匹配方法研究

扫码查看
针对铝型材挤压模设计后图纸重用困难、寻找相似型材面模具效率低等问题,提出基于型材面结构特征的匹配算法,根据新型材面图像匹配数据库中相似的型材面图像,寻找对应的模具结构用于参考.该算法利用YOLOv5算法有效识别型材面结构特征,自动标注结构位置,通过计算对应结构在型材面中的位置、面积与类型得出结构相似度,并采用最小费用最大流算法对结构进行一对一匹配,以获得最佳匹配结果.
Research on matching method based on structural characteristics of profile surface
In view of the difficulties in drawing reuse after aluminum extrusion die design and the low efficiency of finding similar profile surface dies,a matching algorithm based on the structure characteristics of profile surface was proposed.According to the new profile surface image match-ing database of similar profile surface images,the corresponding die design was found for refer-ence.It used YOLOv5 algorithm to effectively identify the structural features of the profile surface and automatically mark the structure position.By calculating the position,acreage,and type of the corresponding structure on the profile surface,the structure similarity was obtained,and the struc-ture was matched one-to-one according to the idea of minimum cost and maximum flow to obtain the best matching result.

aluminum profile sectionstructural characteristicsimage matchingmachine learning

周钢亮、秦昊、林卓胜、张昱、王培龙

展开 >

五邑大学 智能制造学部,广东 江门 529020

广东省科学院智能制造研究所,广东 广州 510075

铝型材截面 结构特征 图像匹配 机器学习

铝型材挤压工艺过程管控系列工业软件研发及应用示范项目

2021B0101220006

2024

模具工业
桂林电器科学研究所

模具工业

影响因子:0.637
ISSN:1001-2168
年,卷(期):2024.50(5)
  • 14