Ling-Feng Shi, Fan Zheng, and Yifan Shi
Indoor location, point–line features, plane features, multi-constraint optimisation algorithm, simultaneous localisation and mapping
Given a large number of repetitive textures or weak texture scenes in indoor environments, simultaneous localisation and mapping (SLAM) systems based on point features are often prone to tracking failures. We propose a multi-constrained optimisation algorithm (MCOA), which is a point–line–plane SLAM combining planar and point–line features. By adding point–line–plane features to feature extraction, matching, pose optimisation, and keyframe selection, MCOA improves its performance in indoor scenes. In addition, in order to improve the operation efficiency and detection accuracy, the ORB method and local feature description are improved. Experimental results show that MCOA into ORB-SLAM2 will significantly improve the system’s robustness in weak texture scenes and positioning accuracy.
Important Links:
Go Back
