Natural landmarks based localization algorithm for indoor robot with binocular vision

Self-localization in unknown environment is one of the most fundamental tasks for mobile robot. In this paper, a novel natural landmarks based localization method is proposed for indoor robot equipped with binocular vision, ceiling corner is taken as natural landmarks. The absolute location of the robot is determined according to the principle of triangular localization. The ceiling corner feature extraction is very importance for localization. Firstly, FAST with adaptive double threshold is proposed to extract feature points, which is described by 32-dimension SIFT descriptor. Then the absolute location is determined by matching feature points with landmarks database based on SIFT. Finally, the feasibility and effectiveness of proposed localization method is demonstrated.

[1]  Trevor Darrell,et al.  DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition , 2013, ICML.

[2]  Zhang Yanli,et al.  The research of video tracking based on improved SIFT algorithm , 2016, 2016 IEEE International Conference on Mechatronics and Automation.

[3]  Tom Drummond,et al.  Faster and Better: A Machine Learning Approach to Corner Detection , 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Long Ye,et al.  An improved SIFT algorithm based on adaptive threshold canny , 2015, 2015 8th International Congress on Image and Signal Processing (CISP).

[5]  Xiangjun Zou,et al.  Localisation of litchi in an unstructured environment using binocular stereo vision , 2016 .

[6]  Xiaorui Zhu,et al.  Vision-Based Semantic Unscented FastSLAM for Indoor Service Robot , 2015 .

[7]  Wei Hong,et al.  Research on binocular vision absolute localization method for indoor robots based on natural landmarks , 2015, 2015 Chinese Automation Congress (CAC).