论文部分内容阅读
提出了基于智能空间的家庭服务机器人同步定位与地图构建(SLAM)方法.利用双目立体视觉传感器提取环境特征,获取环境中物体的Harris角点,通过立体匹配算法获取角点的三维几何信息,同时获取环境中这些几何特征对应的图像特征信息,并将混合信息进行绑定,作为实时更新信息存入智能空间信息库中,构建出三维立体混合特征地图.在SLAM实现过程中,首先建立系统模型并对该模型进行重构以实现线性化;其次移动机器人与智能空间实时地进行交互,实现快速数据关联;最后利用卡尔曼滤波算法处理信息的不确定性,估计出机器人的位姿,同时保存环境特征,逐步构建出环境地图.实验表明,该方法实时性好、精确度高.
This paper proposes a synchronization and mapping (SLAM) method for home service robot based on smart space.Using binocular stereo vision sensor to extract environmental features, Harris corners of objects in environment are obtained, and 3D geometry information of corner points is obtained through stereo matching algorithm, At the same time, the image feature information corresponding to these geometric features in the environment is acquired, and the mixed information is bound as a real-time updated information into the intelligent spatial information database to construct a three-dimensional mixed feature map. In the SLAM implementation process, firstly, a system Model and reconstruct the model to realize linearization. Secondly, the mobile robot interacts with the smart space in real time to realize fast data association. Finally, the Kalman filter algorithm is used to process the uncertainty of the information to estimate the pose of the robot, meanwhile, Save the environmental features, and gradually build the environment map.Experiments show that this method has good real-time and high accuracy.