HU moments aren't invariant for scaling in the discrete state, so they are improved in the paper. The improved moments are consistent with region, boundary and discrete situation. Therefore, they are applied to three-dimensional object recognition. Firstly the improved moments are calculated. Then the similarity measure is computed between objects to be recognized and the standard one. Finally experiments are simulated by MATLAB, and experimental results demonstrate that the improved moments are invariant to the translation, rotating and scaling of objects, the recognition rate is relatively high and the proposed algorithm has some practical value. So the feasibility of the proposed method is proved in the paper.
[1]
Ying Huang,et al.
Recognition of 3-D objects in multiple statuses based on Markov random field models
,
2004,
2004 IEEE International Conference on Acoustics, Speech, and Signal Processing.
[2]
Ming-Kuei Hu,et al.
Visual pattern recognition by moment invariants
,
1962,
IRE Trans. Inf. Theory.
[3]
Mingyuan Zhao,et al.
A method for 3-d object recognition based on support vector machine: A method for 3-d object recognition based on support vector machine
,
2009
.
[4]
Chaur-Chin Chen.
Improved moment invariants for shape discrimination
,
1993,
Pattern Recognit..