基于Kinect传感器的移动机器人室内三维环境创建

Kinect-based 3D indoor environment building for mobile robot

针对移动服务机器人需要感知室内工作环境问题,提出了一种基于低成本Kinect传感器的三维环境创建实用方法。机器人在移动过程中,Kinect传感器实时采集RGB信息和深度信息,将RGB图像信息映射到深度图像信息中,采用联合双边滤波对深度图像进行预处理,获得质量比较高的点云数据。对采集到的大量点云数据,采用半径滤波器算法对点云进行精简,剔除离群点,减少点云数量,提高匹配速度。点云配准时,采用NDT算法完成初始配准,从而得到图像帧间粗略的转换关系,并运用GICP算法对采集的多视角点云数据进行精确配准,得到拼接的三维点云场景。实验结果表明:本文方法具有更好的重建效果和更高的效率,可以应用于室内场景三维环境创建。

In order to solve the problem of indoor environment self-perception for mobile service robots,a new 3 D environment reconstruction method based on low cost Kinect sensor is proposed.Kinect sensor can collect RGB information and depth information in course of robot moving. Firstly,the RGB image information is mapped to the depth image information; and then,the Joint Bilateral filtering is used to preprocess the depth image to obtain the high quality point cloud data. Secondly,the radius filtering algorithm is used for the simplification of the collected large point cloud data to eliminate outliers,reduce the number of cloud points,and improve the matching speed. Thirdly,the NDT（ Normal Distributions Transform） algorithm is adopted to complete the initial registration,and the rough transition relation between frames is obtained. Finally,the GICP（ GeneralizedIterative Closest Point） algorithm is used to register the multi view point cloud data accurately and the reconstruction 3 D scenes of indoor environment are obtained. The experimental results show that the proposed method has better reconstruction effect and higher efficiency,and can be applied to 3 D scene construction of indoor environment.