基于图形路标特征的精确定位研究

Research on Precise Location Based on Graphic Landmark Features

定位系统是智能移动机器人中一项重要的基础性工作,精确定位对机器人的实时导航具有重要的意义。随着计算机图像处理技术的发展,基于图像的定位技术也在不断地进步。该文提出一种基于图形路标的精确定位算法。首先对采集图像进行数据压缩处理,去除图像中冗余的信息,减少计算量、提高运算效率;其次通过SIFT算法对不同尺度空间下的图像进行特征提取,使得不同背景下的图形路标依然能够匹配;最后通过三角定位法来实现移动机器人的精确定位。实验结果表明,该算法是移动机器人实现精确定位和自主导航的一种可行方法。

Position system is an important and foundational part of intelligent mobile robots, precise position is indispensable to real-time navigation of robots. With the development of computer image processing technology,image-based position technology is improving constantly. In this paper, we proposed a precise position algorithm based on pattern signpost. Firstly, the data of image is compressed to remove the redundant information, reduce the computational complexity, and improve the operation efficiency. Then we use scale invariant feature transform(SIFT) algorithm to extract the feature of image in different scale spaces. Finally we apply the triangular positioning method to achieve the precise positioning of moving robots. The experiment result proves that the proposed method is feasible for fast and accurate autonomous navigation of moving robots.