Recognition and Classification for Vision Navigation Vehicle in Agricultural Environment based on MSBN
To solve the uncertain problem in the agricultural environment recognition and classification for vehicles, an environment recognition algorithm for vehicles based on inference in the multiply sectioned Bayesian network (MSBN) is proposed. This method represents multiple image sensor systems into sub-Bayesian networks in the MSBN. With the existing local and global exact inference algorithm in MSBN, the presented method can improve the recognition performance by fusion multi-source partial observation evidences from sub-Bayesian networks via their effective updated belief communication among the subnets. Experimental results illustrate that this MSBN-based agricultural environment recognition and classification approach for vehicles’ navigation system can provide more accurate results than the existing Bayesian network method, with the attractive handling with uncertain and incomplete observation in the single sensor system.
Recognition and Classification Agricultural Environment Multiply Sectioned Bayesian Network (MSBN) Inference
Wenqiang Guo Zoe Zhu Yongyan Hou Ju Fu
School of Electrical and Information Engineering, Shaanxi Univ. of Sci. and Tech., Xian, China, 710 Department of Computing Information Science, University of Guelph,Canada,N1G 2W1 School of Electrical and Information Engineering, Shaanxi Univ. of Sci. and Tech., Xian, China ,710
国际会议
The 24th Chinese Control and Decision Conference (第24届中国控制与决策学术年会 2012 CCDC)
太原
英文
1991-1995
2012-05-23（万方平台首次上网日期，不代表论文的发表时间）