dataset_env_WALL10_task_crosswallpickcube_numepi_20_ctrl_cartesian

该数据集是一个机器人学习数据集，使用LeRobot工具创建。数据集包含20个训练片段，总帧数为3560，数据文件大小为100MB，视频文件大小为200MB，帧率为15fps。机器人类型为panda，特征包括来自主要摄像头和手腕摄像头的视频观察（分辨率256x256，3通道，30fps），以及状态观察（如笛卡尔坐标、关节角度、目标位置等）和动作数据（7维浮点数）。数据以parquet格式存储，视频以mp4格式存储，采用av1编码。数据集适用于机器人控制任务，如跨墙拾取立方体等。This is a robotics learning dataset developed using the LeRobot toolkit. It contains 20 training segments with a total of 3560 frames. The data file has a size of 100 MB, while the total size of the video files is 200 MB, with a frame rate of 15 fps. The robot platform used is the Panda robot. The dataset includes video observations from the main camera and wrist camera (resolution: 256x256, 3-channel, 30 fps), state observations such as Cartesian coordinates, joint angles, target positions and more, as well as 7-dimensional floating-point action data. The dataset is stored in Parquet format, and the videos are saved in MP4 format with AV1 encoding. This dataset is applicable to robotics control tasks such as picking up cubes across walls.

原始信息汇总

数据集概述

• 名称: cbrian/dataset_env_WALL10_task_crosswallpickcube_numepi_20_ctrl_cartesian
• 许可证: Apache-2.0
• 任务类别: 机器人学 (Robotics)
• 数据集创建工具: LeRobot
• 可视化链接: 可视化该数据集

数据集结构

• 机器人类型: Panda (panda)
• 总片段数: 20
• 总帧数: 3560
• 总任务数: 1
• 数据分块: 每块1000帧 (chunks_size: 1000)
• 数据文件大小: 100 MB
• 视频文件大小: 200 MB
• 帧率: 15 FPS
• 数据集划分: 全部20个片段用于训练 (train: "0:20")
• 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

特征字段

特征名称	数据类型	形状	说明
observation.images.primary	video	[256, 256, 3]	主摄像头图像 (高度、宽度、通道)，AV1编码，30 FPS，无音频
observation.images.wrist	video	[256, 256, 3]	腕部摄像头图像 (高度、宽度、通道)，AV1编码，30 FPS，无音频
observation.state.cartesian	float32	[6]	笛卡尔坐标状态 (x, y, z, roll, pitch, yaw)
observation.state.gripper	float32	[1]	夹爪状态 (gripper)
observation.state.joints	float32	[7]	关节状态 (joint_0 至 joint_6)
observation.state.target	float32	[6]	目标状态 (target_x, target_y, target_z, target_roll, target_pitch, target_yaw)
observation.state	float32	[20]	组合状态 (包含笛卡尔坐标、夹爪、关节和目标状态的全部20个元素)
action	float32	[7]	动作 (x, y, z, roll, pitch, yaw, gripper)
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

备注

• 数据集不包含深度图 (video.is_depth_map: false)
• 无音频信息 (has_audio: false)
• 引用信息暂缺 ([More Information Needed])