JHU CoSTAR Block Stacking Dataset

JHU CoSTAR Block Stacking Dataset 包括一个真实的机器人，它试图在具有挑战性的照明和必须避免的可移动垃圾箱障碍物的场景中将彩色儿童积木堆叠近 12,000 次。首先查看下面的示例，图像按时间从左到右排序。请注意截然不同的照明条件、毛绒玩具干扰物的存在、3 或 4 个积木的堆叠、物体磨损以及各种垃圾箱位置。

The JHU CoSTAR Block Stacking Dataset features a real robotic system that carries out nearly 12,000 stacking attempts with colored children's building blocks. These trials take place in scenes with challenging lighting conditions and movable bin obstacles that must be avoided. First review the examples below, where images are ordered chronologically from left to right. Please note the distinct lighting conditions, presence of plush toy distractions, stacks composed of 3 or 4 blocks, signs of object wear and tear, and various bin placements.

搜集汇总

背景与挑战

背景概述

JHU CoSTAR Block Stacking Dataset是一个包含真实机器人操作的数据集，记录了近12,000次在具有挑战性照明和可移动障碍物场景下的彩色儿童积木堆叠过程。数据集特点包括不同的照明条件、干扰物的存在、积木堆叠的多样性以及物体磨损等。