基于稀疏表示的多类融合样本中特定目标识别  

Special object recognition based on sparse representation in multiclass fusion sample

作　　者：查长军[1,2] 韦穗[1] 杨海蓉[3] 丁大为[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,合肥230039 [2]合肥学院机器视觉与智能控制技术重点实验室,合肥230601 [3]合肥师范学院数学系,合肥230061

出　　处：《吉林大学学报（工学版）》2014年第3期822-827,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：NSFC-广东联合基金项目(U1201255);高等学校博士学科点专项科研基金项目(20113401130001);国家自然科学青年基金项目(61201227);安徽省自然科学基金青年项目(1208085QF114);安徽省高校省级自然科学一般项目(KJ2011B131;KJ2013B224)

摘　　要：针对WSN及轮廓检测系统的特点,给出了一种基于WSN的数据处理方法。该方法首先通过主分量分析提取样本特征,然后采用叠加方式对不同样本的特征进行融合,给出数学模型;并以此模型为基础,提出一种新的基于稀疏表示的多类融合样本中特定目标识别算法,该算法根据超完备字典下主要非零系数的分布情况识别出特定目标;数值仿真与实验结果验证了本文算法的有效性,综合性能优于传统方法。According to the characteristics of Wireless Sensor Network （WSN） and the profile detecting system, a data processing method based on WSN is proposed. In this method, first, the sample features are extracted by principal component analysis; then, the features of different samples were fused using accumulate mode. A mathematical model was given and on the basis of this model, a novel algorithm of special object recognition based on sparse representation in multiclass fusion sample was proposed. This algorithm recognizes the special target according to distribution of the main non-zero coefficients under an over-complete dictionary. Numerical simulation and experimental results demonstrate the effectiveness of the proposed algorithm, the comprehensive performance is better than the traditional methods.

关 键 词：信息处理技术 无线传感器网络 稀疏表示 轮廓识别 无人值守地面传感器 

分 类 号：TN911.74[电子电信—通信与信息系统]