基于移动机器人听觉–视觉的目标人定位

Human Localization Based on Audio-Video System on Mobile Robot

摘要: 介绍了一种用于移动机器人搜索目标人方位的听觉和视觉一体的定位方法。针对目标人的位置，利用机器人听觉定位系统进行声源水平方向初定位。在此基础上，用视觉检测人脸中心在摄像头视野中的位置，进行水平角度补偿，引导机器人走向目标人。本文的创新点在于提出了一种四元十字麦克风阵和单目视觉联合对目标人进行定位的方法。离线实验表明，这种声视一体的定位方法在家用机器人目标人定位上是可行的。

Abstract: In this paper, we introduce an audio-video system to help mobile robot to locate human. For the position of human, the auditory system of robot is used to locate the sound azimuth. Then, the survey uses video system to detect the abscissa of human face in the video for horizontal angle compensation; the robot is then guided to the user. The in- novation of this paper is to put forward a kind of method to locate target people using a quaternary cross microphone array and monocular vision. The off-line results show that the audio-video system is feasible to locate human in the in- telligent information household robot.