Delta-Gear/UR10_Random_Pose

UR10随机生成姿态数据集是通过NVDIA Isaac Sim模拟UR10机械臂生成的。所有样本均随机化生成，包括可行的6自由度姿态（无自碰撞或环境碰撞）以及相机的位置、方向和距离。部分数据集包含域随机化，如光照和纹理的随机化。每个样本为1080x1080像素的RGB图像，并附带包含68个标签的robot_state.csv文件，涵盖目标关节角度、实际关节角度、位置、旋转等信息。数据集包含6个子集，分别模拟不同的环境（如仓库、医院、房间等），图像数量从1000到5000不等。

The UR10 Randomly Generated Pose Dataset is generated by simulating the UR10 Robotic Manipulator Arm using NVDIA Isaac Sim. All samples are randomized, including viable 6DoF poses (without collision with self or environment) and camera position, orientation, and distance. Some sets include domain randomization, such as randomized lighting and textures. Each sample is a 1080x1080px RGB image accompanied by a robot_state.csv file with 68 labels, including target joint angles, actual joint angles, positions, rotations, etc. The dataset consists of 6 subsets, each simulating different environments (e.g., warehouse, hospital, room) with image counts ranging from 1000 to 5000.