每天一分钟,带你读遍机器人顶级会议文章
标题:Probabilistic Data Association for Semantic SLAM
作者:Sean L. Bowman, Nikolay Atanasov, Kostas Daniilidis, George J. Pappas
来源:ICRA 2017
播音员:刘畅
编译:赵博欣
欢迎个人转发朋友圈;其他机构或自媒体如需转载,后台留言申请授权
摘要
今天介绍的文章是“Probabilistic Data Association for Semantic SLAM ”——面向语义SLAM的概率数据关联,该文章发表在ICRA2017。
传统的SLAM算法都依赖于低级别的几何特征,比如点、线和平面。这些特征无法对环境中观测到的标志物进行语义标识。而且,基于这些低级特征使得回环检测通常依赖于摄像机的视角,而且在模糊或重复性的纹理环境中容易检测失败。另一方面,通过目标识别可以推测出标志物种类的大小,从而产生一小组易于识别的标志物,非常适用于与视角无关的闭环检测。然而,当地图中存在多个同类物体时,则需要对关键的数据进行关联。但数据关联和识别通常是用离散方法解决的离散问题,而传统SLAM是一个对尺度信息的连续优化问题。在本文中,我们将传感器状态和语义标志物的位置信息建模成一个优化问题,融合了尺度信息,语义信息和数据关联。然后把它分解为两个相互关联的问题:一个是离散数据关联和标志物种类概率估计,另一个是对尺度状态的连续优化。估计出的标志物和机器人姿态影响着数据的关联和标志物种类的分布,而这反过来又影响机器人-标志物姿态的优化。最后,通过室内和室外数据集验证了本文算法的性能。
下图展示了对关键帧图像进行ORB特征检测(绿色点)和目标检测的情况
下图显示的是算法估计得到的传感器轨迹(蓝色)和标志物的位置与种类。论文中附带的视频显示该定位过程可以实时处理。
下图显示的是在室内办公室环境下不同算法的定位轨迹图对比
Abstract
Traditional approaches to simultaneous localization and mapping (SLAM) rely on low-level geometric features such as points, lines, and planes. They are unable to assign semantic labels to landmarks observed in the environment. Furthermore, loop closure recognition based on low-level features is often viewpoint-dependent and subject to failure in ambiguous or repetitive environments. On the other hand, object recognition methods can infer landmark classes and scales, resulting in a small set of easily recognizable landmarks, ideal for view-independent unambiguous loop closure. In a map with several objects of the same class, however, a crucial data association problem exists. While data association and recognition are discrete problems usually solved using discrete inference, classical SLAM is a continuous optimization over metric information. In this paper, we formulate an optimization problem over sensor states and semantic landmark positions that integrates metric information, semantic information, and data associations, and decompose it into two interconnected problems: an estimation of discrete data association and landmark class probabilities, and a continuous optimization over the metric states. The estimated landmark and robot poses affect the association and class distributions, which in turn affect the robot-landmark pose optimization. The performance of our algorithm is demonstrated on indoor and outdoor datasets.
如果你对本文感兴趣,想要下载完整文章进行阅读,可以关注【泡泡机器人SLAM】公众号。
回复关键字“20180127”,即可获取本文下载链接。
泡泡机器人SLAM的原创内容均由泡泡机器人的成员花费大量心血制作而成,希望大家珍惜我们的劳动成果,转载请务必注明出自【泡泡机器人SLAM】微信公众号,否则侵权必究!同时,我们也欢迎各位转载到自己的朋友圈,让更多的人能进入到SLAM这个领域中,让我们共同为推进中国的SLAM事业而努力!
商业合作及转载请联系liufuqiang_robot@hotmail.com