HAMMER
收藏资源简介:
HAMMER数据集是一个高度精确的多模态室内深度估计数据集,由慕尼黑工业大学和华为诺亚方舟实验室共同创建。该数据集包含13,000帧图像,涵盖了多种传感器数据,包括飞行时间(ToF)、立体视觉、主动立体视觉以及单目RGB+P数据。数据集通过3D扫描和渲染技术生成高可靠性的深度图,旨在为深度估计和传感器融合方法提供一个可靠的基础。HAMMER数据集适用于研究室内场景中的深度估计问题,特别是在处理具有挑战性的日常场景内容时,如纹理缺失区域、反射材料和透明材料等。
The HAMMER dataset is a highly accurate multi-modal indoor depth estimation dataset jointly created by the Technical University of Munich and Huawei Noah's Ark Lab. This dataset contains 13,000 image frames, covering multiple types of sensor data including Time-of-Flight (ToF), stereo vision, active stereo vision, and monocular RGB+P data. The dataset generates high-reliability depth maps through 3D scanning and rendering technologies, aiming to provide a reliable foundation for depth estimation and sensor fusion methods. The HAMMER dataset is suitable for researching depth estimation problems in indoor scenes, especially when dealing with challenging daily scene contents such as texture-deficient areas, reflective materials, and transparent materials.




