一种基于模拟退火的特征层融合模式识别实现方法  

Study of feature level fusion pattern recognition based on simulated annealing

作　　者：胡勇[1] 高隽[1] 柴斌[1] 胡良梅[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2004年第6期587-591,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(60175011);安徽省自然科学基金资助项目(01042301);优秀青年科技基金资助项目(04042044);安徽省重点科研基金资助项目(03021012)

摘　　要：文章提出一种特征层融合模式识别的方法,定义"融合特征"和"特征融合系数"实现模式的特征融合。使用模拟退火算法搜索最佳的"特征融合系数",从而提高模式识别的正确性,降低分类决策的模糊性。特征层融合的交通标志形状识别实验和数字字符识别实验均表明:该方法具有以上两方面的优点。In this paper, a method of feature level fusion pattern recognition(FLFPR) is presented based on simulated annealing. In this new approach, two parameters, Fusion Feature and Feature Fusion Coefficient are defined to fuse the features of multi-sensor from a pattern. The simulated annealing algorithm is applied to search the optimal Feature Fusion Coefficient. Therefore the correctness of pattern recognition system can be improved greatly,and the fuzziness of decision on classification and recognition reduced effectively. The shape recognition experiment and number recognition experiment of traffic signs show that the presented method is feasible.

关 键 词：信息融合 模拟退火 模式识别 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]