Tomato Truss Grasping Dataset

该数据集包含了25种不同的番茄串，用于评估在不同环境下的抓取方法，包括无障碍和有障碍的环境。此外，数据集还包括了采用不同抓取策略的试验，并记录了不同抓取方法的成功与失败率。该数据集的规模涉及25种番茄串在多种场景下的测试，其任务是对使用机器人系统进行番茄串的抓取与操作研究。This dataset comprises 25 distinct tomato trusses, designed for evaluating grasping methods across a range of environments, including both unobstructed and obstructed scenarios. Additionally, the dataset includes trials that employ different grasping strategies, and documents the success and failure rates of various grasping methods. The dataset involves testing the 25 tomato trusses across multiple scenarios, and its core purpose is to support research on grasping and manipulation of tomato trusses using robotic systems.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于机器人抓取藤蔓番茄的研究项目，基于硕士论文“Learning Approach To Robotic Grasping Of Vine Tomatoes”开发。它包含代码和配置，支持Franka Emika Panda机器人、ROS Noetic环境，并集成点云处理和深度学习网络（如PyTorch）以实现物体检测和抓取任务。