基于Kinect传感器的机器人室内环境检测方法

针对移动机器人室内环境检测问题,提出了一种基于Kinect传感器的目标物体检测方法.利用Kinect传感器采集的视频图像和深度数据来实现对机器人工作环境中已知特征目标物体和完全未知目标物体的检测及定位.对于已知特征目标通过颜色特征分析来完成检测,而对于完全未知的物体则通过深度地面消除算法和提取深度图像的轮廓来进行检测.利用传感器成像模型对检测出的目标区域进行三维空间定位,从而获取目标物相对于机器人的空间位置信息.基于移动机器人平台进行实验,结果表明,该方法能够有效地实现室内环境信息的检测及定位.

Indoor environment detection method of robot based on Kinect sensor

Aiming at the indoor environment detection problem of mobile robot, a target object detection method based on Kinect sensor was proposed. With the video image and depth data captured by Kinect sensor, the detection and localization of target objects with the known features and the completely unknown target objects in the work environment of robot were realized. For the target objects with the known features, the detection could be implemented with the color feature analysis. For the completely unknown objects, the detection could be implemented through adopting the depth ground elimination algorithm and extracting the contour of depth image. Furthermore, the three dimensional spatial localization for the detected object area was performeed with the sensor imaging model, and thus the relative spatial location information between the subject object and robot was obtained. The experiments were implemented based on a mobile robot platform. The results show the proposed method can effectively achieve the detection and localization of indoor environment information.