遇见数据集

test_kovla_ur5_sample

收藏
Hugging Face2026-09-11 更新2026-09-11 收录
官方服务:

资源简介:

该数据集使用LeRobot创建,是一个针对UR5机械臂的机器人操控数据集。包含来自两个摄像头(egoview和wristcam)的640x480 RGB视频流(30fps,H.264编码),以及7维关节位置动作和状态观测(包括6个关节角度和夹爪位置)。此外,还包含任务相关的注释,如分割ID、初始姿态位置和网格坐标等。数据共1个episode,1360帧,1个任务,适合用于模仿学习和机器人操作研究。

This dataset is created with LeRobot and is a robot manipulation dataset for the UR5 arm. It contains 640x480 RGB video streams (30 fps, H.264 coded) from two cameras (egoview and wristcam), as well as 7-dimensional joint position action and state observations (including 6 joint angles and gripper position). Additionally, task-related annotations such as segmentation IDs, initial pose positions, and grid coordinates are included. The dataset consists of 1 episode, 1360 frames, and 1 task, suitable for imitation learning and robot manipulation research.

提供机构:
guunman
创建时间:
2026-09-11
原始信息汇总

数据集概述

基本信息

  • 数据集地址:https://huggingface.co/datasets/guunman/test_kovla_ur5_sample
  • 许可证:apache-2.0
  • 任务类别:robotics
  • 标签:LeRobot
  • 创建工具:LeRobot(https://github.com/huggingface/lerobot)
  • 机器人类型:ur5

数据规模

  • 总回合数(total_episodes):1
  • 总帧数(total_frames):1360
  • 总任务数(total_tasks):1
  • 帧率(fps):30
  • 代码库版本(codebase_version):v3.0
  • 划分方式(splits):train: 0:1
  • chunks_size:1000
  • data_files_size_in_mb:100
  • video_files_size_in_mb:200

数据文件配置

  • 配置名称:default
  • 数据文件:data//.parquet
  • 数据路径(data_path):data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
  • 视频路径(video_path):videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4

特征结构

action

  • 数据类型:float32
  • 形状:7
  • 字段名称:joint1.pos, joint2.pos, joint3.pos, joint4.pos, joint5.pos, joint6.pos, gripper.pos

observation.state

  • 数据类型:float32
  • 形状:7
  • 字段名称:joint1.pos, joint2.pos, joint3.pos, joint4.pos, joint5.pos, joint6.pos, gripper.pos

observation.images.egoview

  • 数据类型:video
  • 形状:480, 640, 3
  • 字段名称:height, width, channels
  • 视频信息
    • is_depth_map:false
    • video.height:480
    • video.width:640
    • video.codec:h264
    • video.pix_fmt:yuv420p
    • video.fps:30
    • video.channels:3
    • has_audio:false
    • video.g:30
    • video.crf:28
    • video.preset:veryfast
    • video.fast_decode:0
    • video.video_backend:pyav
    • video.extra_options:{}

observation.images.wristcam

  • 数据类型:video
  • 形状:480, 640, 3
  • 字段名称:height, width, channels
  • 视频信息
    • is_depth_map:false
    • video.height:480
    • video.width:640
    • video.codec:h264
    • video.pix_fmt:yuv420p
    • video.fps:30
    • video.channels:3
    • has_audio:false
    • video.g:30
    • video.crf:28
    • video.preset:veryfast
    • video.fast_decode:0
    • video.video_backend:pyav
    • video.extra_options:{}

annotation.segment_id

  • 数据类型:int64
  • 形状:1
  • 字段名称:id

annotation.reset.initial_pose.position_id

  • 数据类型:int64
  • 形状:5
  • 字段名称:pen_holder, marble_tray.corner_1, marble_tray.corner_2, marble_tray.corner_3, marble_tray.corner_4

annotation.reset.initial_pose.grid_col

  • 数据类型:int64
  • 形状:5
  • 字段名称:pen_holder, marble_tray.corner_1, marble_tray.corner_2, marble_tray.corner_3, marble_tray.corner_4

annotation.reset.initial_pose.grid_row

  • 数据类型:int64
  • 形状:5
  • 字段名称:pen_holder, marble_tray.corner_1, marble_tray.corner_2, marble_tray.corner_3, marble_tray.corner_4

annotation.reset.initial_pose.x_norm

  • 数据类型:float32
  • 形状:5
  • 字段名称:pen_holder, marble_tray.corner_1, marble_tray.corner_2, marble_tray.corner_3, marble_tray.corner_4

annotation.reset.initial_pose.y_norm

  • 数据类型:float32
  • 形状:5
  • 字段名称:pen_holder, marble_tray.corner_1, marble_tray.corner_2, marble_tray.corner_3, marble_tray.corner_4

timestamp

  • 数据类型:float32
  • 形状:1
  • 字段名称:null

frame_index

  • 数据类型:int64
  • 形状:1
  • 字段名称:null

episode_index

  • 数据类型:int64
  • 形状:1
  • 字段名称:null

index

  • 数据类型:int64
  • 形状:1
  • 字段名称:null

task_index

  • 数据类型:int64
  • 形状:1
  • 字段名称:null

元数据信息

  • 元数据文件:meta/info.json

可视化

可通过 LeRobot 可视化空间查看该数据集:https://huggingface.co/spaces/lerobot/visualize_dataset?path=guunman/test_kovla_ur5_sample

引用信息

  • BibTeX:[More Information Needed]
  • Homepage:[More Information Needed]
  • Paper:[More Information Needed]
二维码
社区交流群
二维码
科研交流群
商业服务