计算机与现代化

• 图像处理 • 上一篇    下一篇

基于网格统计模型的三维目标识别

  

  1. 西北工业大学航天学院,陕西西安710072
  • 收稿日期:2013-12-31 出版日期:2014-04-17 发布日期:2014-04-23
  • 作者简介:作者简介:方雪(1989),女,辽宁抚顺人,西北工业大学航天学院硕士研究生,研究方向:图像处理,模式识别;余瑞星(1978),女,湖北武汉人,副教授,博士,研究方向:图像处理,目标识别和跟踪。
  • 基金资助:
     
    基金项目:国家自然科学基金资助项目(61101191); 航空基金资助项目(20130153003); 上海航天科技创新基金资助项目(SAST201342)

 Gridbased Statistical Model for 3D Object Recognition

  1. School of Astronautics, Northwestern Polytechnical University, Xi’an 710072, China
  • Received:2013-12-31 Online:2014-04-17 Published:2014-04-23

摘要:  

摘要: 提出一种基于网格的统计模型的三维目标识别算法。首先将网格结构引入多视点图像,并针对网格位置,利用三维目标多视点间的关联性,再根据目标的局部不变特征建立统计模型;其次对图像数据库COIL中三维目标的自由度进行扩充;最后在此基础上,对算法的识别性能进行测试。实验结果表明,该算法不仅能有效识别三维目标的类别,而且能够对目标的姿态做出可靠的判断,具有较强的鲁棒性。

关键词: 三维目标识别, 多视点, 网格结构, 目标姿态识别

Abstract:  

 Abstract:  A 3D object recognition algorithm based on statistical model of grid structure is presented in this paper. First, a grid structure is introduced into the multiview points image. Then according to the positions of the grids, the relationship between the multiview points of the 3D object and the local invariant features of the object, the statistical model is established. Second, the degrees of freedom of the 3D object, which is from the image databaseCOIL, are expanded. At last, based on the previous work, the recognition performance of this algorithm is tested. The empirical results illustrate that this 3D object recognition algorithm, with relatively strong robustness, can not only recognize the classification of 3D objects effectively, but also make reliable judgments on the attitudes of the objects.

Key words:

中图分类号: