摘要
根据NMI特征和Jan Flusser提出的仿射不变矩的特点及各自的适用条件,提出了一种组合不变矩。利用所提取的组合不变矩的特征向量,实现了对空中目标的K-mean聚类分析识别。在此基础上将模拟退火机制引入其中,以克服K-mean聚类的局限性和对初始聚类中心的敏感性,最后提出了基于模拟退火的改进K-mean聚类算法。仿真实验结果表明,该方法具有较高的识别精度和较好的抗噪性。
A new combined invariant moment is presented in this paper according to the characteristics of NMI and affine invariant moments proposed by Jan Flusser with their application conditions. Combined invariant moments are extracted as the feature vector, the air target classification based on K -means clustering analysis is implemented successfully. On the basis of systematic analysis of current algorithms, simulated annealing mechanism is inducted into K-means clustering to solve its locality and the sensitive- ness of the initial condition. Then, this paper proposes the improved K -means clustering based on simulated annealing algorithm. The simulation test results show that the proposed method achieves good classification and antinoise performance.
出处
《航空兵器》
2008年第6期28-31,39,共5页
Aero Weaponry
基金
火力控制技术国防科技重点实验室基金资助项目(05C52007)
关键词
仿射不变矩
组合不变矩
K—means聚类
模拟退火算法
目标识别
affine invariant moment
combined invariant moment
K-means clustering
simulated annealing algorithm
target recognition