ALLO
收藏资源简介:
ALLO数据集是由多伦多大学航空航天研究所创建的,用于月球轨道机器人近距离操作中的异常检测。该数据集包含94,890张无异常图像和17,527张有异常图像,每张图像都带有像素级地面真值图。数据集通过Blender的Cycles渲染引擎生成,模拟了月球轨道中的复杂光照条件。创建过程包括模拟月球、地球和太阳的位置,以及机器人手臂的相机视角。ALLO数据集主要用于评估和改进空间机器人系统的视觉异常检测算法,旨在提高月球轨道操作中的自主性和安全性。
The ALLO Dataset was developed by the Aerospace Institute of the University of Toronto for anomaly detection during close-range operations of lunar orbit robots. This dataset includes 94,890 anomaly-free images and 17,527 anomalous images, each paired with a pixel-level ground truth map. The dataset was generated using Blender's Cycles rendering engine, which simulates the complex lighting conditions in lunar orbits. Its creation process simulates the positions of the Moon, Earth, and Sun, as well as the camera viewpoints of the robotic arm. The ALLO Dataset is primarily used to evaluate and improve visual anomaly detection algorithms for space robotic systems, aiming to enhance autonomy and safety in lunar orbit operations.

- 1ALLO: A Photorealistic Dataset and Data Generation Pipeline for Anomaly Detection During Robotic Proximity Operations in Lunar Orbit多伦多大学航空航天研究所 · 2024年



