基于随机森林的三维工件点云分类研究
Research on 3D Workpiece Point Cloud Classification Based on Random Forest

针对三维工件点云分类问题，本文研究了一种基于随机森林的三维工件点云分类方法。首先，将由双目相机获得的工件深度图像和彩色图像转换为彩色工件点云；然后，通过三维点云预处理技术对工件点云进行预处理，从而获得目标工件点云，并提取工件点云特征；最后，使用随机森林模型学习工件点云特征，并进行工件点云分类。通过在T-LESS数据集上的实验，验证了本文研究工作的有效性。在10个测试集场景上的三维工件点云分类中，取得了较优的分类结果。
Aiming at the 3D workpiece point cloud classification problem, this paper studies a 3D workpiece point cloud classification method based on random forest. Firstly, the depth images and color images of the workpieces obtained by the binocular camera are converted into workpiece point clouds; Subsequently, the workpiece point clouds are preprocessed by 3D point cloud preprocessing technology to obtain the target workpiece point clouds and extract the workpiece point cloud features; Finally, the random forest model is used to learn the workpiece point cloud features and perform the classification of the workpiece point cloud. The effectiveness of the research work in this paper is verified by experiments on the T-LESS dataset. In the 3D workpiece point cloud classification on 10 test set scenarios, superior classification results are achieved.