Phocal

Phocal数据集是由慕尼黑工业大学创建的，专注于解决光度挑战性对象（如无纹理、反射表面或透明物体）的3D视觉任务。该数据集通过集成多模态传感器（包括主动立体、I-ToF、D-ToF和RGB-P传感器）和精确的机器人正向运动学，实现了高精度的数据采集。创建过程中，利用机器人和自由手两种方法进行数据标注，确保了数据的高质量和广泛的视角覆盖。Phocal数据集主要应用于3D视觉研究，特别是在处理复杂光照条件下的物体识别和姿态估计问题。

The Phocal Dataset was developed by the Technical University of Munich, focusing on addressing 3D vision tasks involving photometrically challenging objects such as texture-less, reflective or transparent objects. This dataset enables high-precision data collection by integrating multimodal sensors including active stereo, Indirect Time-of-Flight (I-ToF), Direct Time-of-Flight (D-ToF) and RGB-P sensors, alongside accurate robot forward kinematics. During its development, data annotation was conducted using two approaches: robot-based and free-hand methods, which ensure the dataset's high quality and comprehensive viewpoint coverage. The Phocal Dataset is primarily utilized in 3D vision research, particularly for tackling object recognition and pose estimation problems under complex lighting conditions.