三维机械零件数据集 （混元3D生成，暂时未做数据清洗）

本项目打算使用GHEZELBASH F的[1]数据集，采用Hunyuan3D2.0根据生成的图片生成.glb三维模型。为作者本科毕设项目的一个小小分支。作者的毕设打算运用大模型构建三维数据集，通过深度学习算法从不同角度提取数字特征，最后通过特征匹配完成对复杂三维零件的精确识别。不过在数据清洗方面暂时遇到了一个问题。

其中解压所有zip文件，则是所有的.glb文件

[1] GHEZELBASH F, ESKANDARI A H, BIDHENDI A J. A Dataset for Mechanical Mechanisms[EB/OL]. arXiv, 2024[2024-09-30]. arXiv:2409.03763 [cs.CV].

This project intends to utilize the dataset published by GHEZELBASH F[1], and employ Hunyuan3D 2.0 to generate .glb 3D models from the generated images. This is a minor branch of the author's undergraduate graduation project. The graduation project aims to construct a 3D dataset using large language models (LLMs), extract digital features from various perspectives via deep learning algorithms, and finally achieve accurate recognition of complex 3D parts through feature matching. However, a temporary problem has been encountered during the data cleaning process.

Extracting all included zip files will yield all the .glb files.

[1] GHEZELBASH F, ESKANDARI A H, BIDHENDI A J. A Dataset for Mechanical Mechanisms[EB/OL]. arXiv, 2024[2024-09-30]. arXiv:2409.03763 [cs.CV].

搜集汇总

背景与挑战

背景概述

该数据集是由混元3D生成的机械零件三维模型集合，主要用于本科毕设项目中的三维零件识别研究。数据集包含多个.glb格式文件，但尚未进行数据清洗。