A novel generalized grey model and its application for ouliers detection of pose estimation

A novel generalized grey model is presented for ouliers detection of pose estimation in this paper. Firstly, the novel model is obtained by comprehensively considering order of accumulation, background value, grey control parameter. Secondly, accumulations in different directions and model parameters are discussed by using the thought of matrix analysis, followed by analysis of the fitted error and singularity of the new model. Finally, a particle swarm algorithm is performed to obtained the optimal parameters of the model. The results of practical numerical examples demonstrate that the proposed model performs better than the traditional grey model.