3-D Object Recognition Via Wing Features

对于机器人视觉和自动检测系统来说，识别3-D物体仍然是一项艰巨的任务，即使所有感兴趣的物体都有计算机模型可用。这笔助学金将支持一名研究生学习使用“机翼特征”，或表面/边缘/表面描述符，从距离数据中识别刚性三维物体。这项工作包括用于图像分割和特征提取、数据库索引和姿势确定的实验软件开发。

Recognition of 3-D objects continues to be a difficult task for robotic vision and automated inspection systems, even when computer models are available for all objects of interest. This grant will support a graduate student studying the use of "wing features," or surface/edge/surface descriptors, for the identification of rigid 3-D objects from range data. The work includes experimental software development for image segmentation and feature extraction, database indexing, and pose determination.