摘 要:随着智能时代的到来,视觉机器人在自然场景中会遇到行人位姿变化、障碍物遮挡等复杂环境,使特征点误匹配。 文章先对Mask R-CNN 算法应用于动态环境下的SLAM 中进行研究。通过深度神经网络优化SLAM 视觉前端,使得神经网络能够对动态物体进行检测并能在很大程度上识别动态特征点,减少了特征点的误匹配,提高了相机位姿估计的准确性。最后与ORB-SLAM2 进行仿真对比,结果表明,该文算法和ORB-SLAM2 算法相比精度提高了96% 以上,能够明显的提高SLAM 算法匹配的正确率。
关键词:Mask R-CNN;动态环境;特征点匹配;视觉SLAM
中图分类号:TP391.41;TP242 文献标识码:A 文章编号:2096-4706(2020)21-0080-04
Research on ORB-SLAM Based on Mask R-CNN in Dynamic Environment
WANG Weiliang
(Shenyang Jianzhu University,Shenyang 110168,China)
Abstract:With the advent of the intelligent era,visual robots will encounter complex environments such as pedestrian pose change and obstacles occlusion in natural scenes,which makes feature points mismatched. The article first studies the Mask R-CNN algorithm applied to SLAM in a dynamic environment. The SLAM vision front end is optimized by the deep neural network,so that the neural network can detect dynamic objects and identify dynamic feature points to a large extent,reduce the mismatch of feature points,and improve the accuracy of camera pose estimation. Finally,a simulation comparison with ORB-SLAM2 shows that the accuracy of this algorithm is improved by more than 96% compared with ORB-SLAM2,which can significantly improve the accuracy of SLAM algorithm matching.
Keywords:Mask R-CNN;dynamic environment;feature point matching;visual SLAM
参考文献:
[1] 张文. 基于多传感器融合的室内机器人自主导航方法研究 [D]. 合肥:中国科学技术大学,2017.
[2] GRISETTI G,STACHNISS C,BURGARD W. Improved Techniques for Grid Mapping With Rao-Blackwellized Particle Filters [J].IEEE Transactions on Robotics,2007,23(1):34-46.
[3] 石杏喜,赵春霞,郭剑辉. 基于PF/CUKF/EKF 的移动机器人SLAM 框架算法 [J]. 电子学报,2009,37(8):1865-1868.
[4] 刘雅彬,陈万米. 融合激光与视觉信息的自主移动机器人的SLAM 研究 [J]. 工业控制计算机,2015,28(9):70-72+74.
[5] 侯荣波,魏武,黄婷,等. 基于ORB-SLAM 的室内机器人定位和三维稠密地图构建 [J]. 计算机应用,2017,37(5):1439-1444.
[6] SIMONYAN K,ZISSERMAN A. Two-stream convolutionalnetworks for action recognition in videos [C]//Proceedings of the 27thInternational Conference on Neural Information Processing Systems.Cambridge:MIT Press,2014:568-576.
作者简介:王伟良(1995.06—),男,汉族,辽宁沈阳人,研究生在读,研究方向:移动机器人控制技术。