改进3D SLAM算法在移动机器人上的应用

The Application of Improved 3D SLAM Algorithm in Mobiel Robot

实现移动机器人的同步定位与建图(SLAM)是实现移动机器人真正意义上自主导航的关键。越来越多的研究者开始关注基于机器视觉的SLAM技术,这是因为它具有丰富的信息和低廉的价格。针对传统的室内环境下的3D SLAM存在实时性差、鲁棒性低等问题,提出了一种改进的RGB-D SLAM算法,利用主成分分析法(PCA)对特征描述子进行降维处理,以加快图像匹配的速度。同时针对RGB-D SLAM过程中平移和旋转较慢这一特征,提出了一种将图像区域分块匹配的方法,提高了特征点匹配的效率,降低了误匹配率。同时限制所划分的区域内的特征点数目,使得提取到的图像特征更均匀。为了克服原始RGB-D SLAM的效率不佳问题,采用了RTAB-MAP来实现RGB-D SLAM。在后端,g2o用于机器人的轨迹和全局地图优化。

The realization of mobile robot synchronization positioning and construction(SLAM)is the key to achieve the true autonomous navigation of mobile robots in the true sense. More and more researchers are beginning to pay attention to SLAM technology based on machine vision,because it has a wealth of information and low price. In order to solve the problem of poor real-time performance and low robustness of 3D SLAM in the traditional indoor environment,an improved RGB-D SLAM algorithm is proposed in this paper. Principal component analysis(PCA)is used to reduce the dimensionality of feature descriptors so that it will speed up image matching. In addition,to solve the problem of slow translation and rotation during the process of RGB-D SLAM,a method of segmenting the image region is proposed,which improves the efficiency of feature point matching and reduces the false matching rate. At the same time,by limiting the number of feature points in the divided area,the extracted image features are more uniform. To ipmrove the inefficiency of the original RGB-D SLAM,RTAB-MAP is used to implement RGB-D SLAM. At the backend,g2o is used for robot trajectory and global map optimization.