论文部分内容阅读
针对家庭或者办公室等室内环境下双足机器人的自定位问题,提出一种通过对世界平面的单应矩阵进行分解来实现自定位的方法。首先设计一种新型AR-mark人工地标,利用颜色信息分割感兴趣区域提取出AR-mark的特征点,然后对特征点与空间3D点进行匹配。通过直接线性变化法计算出世界平面的单应矩阵后,利用世界平面的单应分解算法获得摄像机坐标系与AR-mark人工地标坐标系间的位置关系,最终实现双足机器人的局部精确定位。实验结果表明,AR-mark人工地标易于识别,信息内容丰富,同时利用世界平面单应矩阵分解算法定位精度高,能够很好地辅助双足机器人完成室内自定位任务。
In order to solve the problem of self-localization of biped robot in indoor environment, such as home or office, a method of self-localization is proposed by decomposing the plane homography matrix in the world. Firstly, a new AR-mark artificial landmark is designed, and the feature points of AR-mark are extracted by using the color information segmentation region of interest, and then the feature points are matched with the 3D points in space. After the homography matrices of the world plane are calculated by the direct linear variation method, the positional relationship between the camera coordinate system and the AR-mark artificial landmark coordinate system is obtained by using the world plane homography decomposition algorithm. Finally, the local accurate positioning of the biped robot is achieved. The experimental results show that AR-mark artificial landmark is easy to recognize and the information content is rich. At the same time, the use of world plane homography matrix decomposition algorithm has high positioning accuracy and can well assist the biped robot to accomplish the indoor self-localization task.