pick-up-green-eraser-place-in-brown-basket-stacked-baskets-rgbd_20260728_181849
收藏资源简介：
该数据集是一个用于机器人操作任务的数据集，具体任务为“拾取绿色橡皮并将其放置在棕色篮子中（篮子堆叠）”。它由LeRobot工具创建，包含24个episodes，总计21532帧数据，采样率为30fps。数据特征包括：动作（action，7维浮点数组，对应机器人关节位置和夹爪位置）、观测状态（observation.state，7维浮点数组，与动作相同）、来自腕部摄像头的RGB图像（480x640x3）和深度图像（480x640x1，深度单位为毫米）、来自顶部摄像头的RGB图像（480x640x3），以及时间戳、帧索引、episode索引等元数据。数据集使用Apache 2.0许可证，机器人类型为“rebot_b601_follower”。数据以parquet格式存储，视频以mp4格式存储，总数据文件大小为100MB，视频文件大小为200MB。
This dataset is designed for robotic manipulation tasks, specifically the task of "picking up a green rubber object and placing it into a stacked brown basket". It was created with the LeRobot tool, and contains 24 episodes totaling 21532 frames at a sampling rate of 30 fps. The dataset includes the following features: action (a 7-dimensional floating-point array corresponding to robot joint positions and gripper position), observation.state (a 7-dimensional floating-point array identical to the action), RGB images (480×640×3) and depth images (480×640×1 with depth unit in millimeters) from the wrist-mounted camera, RGB images (480×640×3) from the top-mounted camera, as well as metadata such as timestamps, frame indices and episode indices. The dataset is licensed under the Apache 2.0 license, and the robot type is "rebot_b601_follower". The dataset's data is stored in Parquet format, while the video files are stored in MP4 format. The total size of the data files is 100 MB, and the total size of the video files is 200 MB.
数据集概述
- 数据集名称: pick-up-green-eraser-place-in-brown-basket-stacked-baskets-rgbd_20260728_181849
- 创建工具: LeRobot (https://github.com/huggingface/lerobot)
- 许可证: Apache-2.0
- 任务类别: 机器人学 (robotics)
数据集结构
- 机器人类型: rebot_b601_follower
- 总片段数 (episodes): 24
- 总帧数: 21,532
- 总任务数: 1
- 帧率 (fps): 30
- 数据文件大小: 约 100 MB
- 视频文件大小: 约 200 MB
- 数据集划分:
- 训练集: 24 个片段 (索引 0-23)
特征说明
| 特征名称 | 数据类型 | 形状 | 描述 |
|---|---|---|---|
action |
float32 | (7,) | 机器人关节动作 (shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_yaw, wrist_roll, gripper) |
observation.state |
float32 | (7,) | 机器人关节状态 (同动作维度) |
observation.images.wrist |
video (AV1) | (480, 640, 3) | 腕部 RGB 摄像头图像 (高度x宽度x通道) |
observation.images.wrist_depth |
video (HEVC) | (480, 640, 1) | 腕部深度图 (单位: mm, 灰度12位LE) |
observation.images.top |
video (AV1) | (480, 640, 3) | 顶部 RGB 摄像头图像 |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 片段索引 |
index |
int64 | (1,) | 索引 |
task_index |
int64 | (1,) | 任务索引 |
数据文件格式
- 数据文件路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4 - chunks 大小: 每块 1000 帧
深度图特殊配置
- 深度编码: HEVC (h265), 无损压缩 (lossless=1)
- 深度范围: 最小 0.01m, 最大 10.0m
- 偏移量 (shift): 3.5
- 使用对数映射 (use_log): true
