基于稀疏直接法的SLAM算法  

SLAM Algorithm Based on Sparse Direct Method

作　　者：张超[1] 曹雄 徐春凤[1] 韩成[1] 白烨[2] ZHANG Chao;CAO Xiong;XU Chunfeng;HAN Cheng;BAI Ye(School of Computer Science and Technology,Changchun University of Science and Technology,Changchun 130022;School of Optical and Electronical Information,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学计算机科学技术学院,长春130022 [2]长春理工大学光电信息学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2018年第5期101-104,共4页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：国家自然基金青年科学基金项目(61602058);吉林省科技攻关计划项目(20170203003GX;20170203004GX)

摘　　要：SLAM技术已经应用到了智能机器人、无人机、AR/VR、无人驾驶等领域中,但当前的SLAM算法在运行速度、相机轨迹精度和鲁棒性等方面依然存在改善空间。为了提高SLAM中相机轨迹精度,提出了基于稀疏直接法的SLAM算法。该算法使用改进的Shi-Tomasi特征点检测算法进行特征点的提取,然后依据提取的特征点采用直接法进行相机位姿估计,同时利用构建的地图进行位姿优化,有效地提高了相机位姿估计精度。针对TUM标准数据集,通过对比分析可知,基于稀疏直接法的SLAM算法可以有效地减少误差,相机轨迹精度优于ORB-SLAM2算法。SLAM technology has been applied to intelligent robots,unmanned aerial vehicles,AR/VR,and driverless fields.However,current SLAM algorithms still have improved space in terms of operating speed,camera track accuracy,and robustness.In order to improve the accuracy of camera trajectories in SLAM,SLAM algorithm based on sparse direct method is proposed.An improved Shi-Tomasi feature point detection algorithm is used to extract feature points;then the direct method is used to perform camera pose estimation based on the extracted feature points;and the constructed map is used for pose optimization to effectively improve the estimated accuracy of camera pose.According to the TUM standard data set,through comparative analysis,the SLAM algorithm based on sparse direct method can effectively reduce the error,and the camera trajectory accuracy is better than the ORB-SLAM2 algorithm.

关 键 词：稀疏特征点 直接法 相机轨迹 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]