摘要
为提高目标识别的准确性和快速性,提出了基于全局和局部特征对目标识别的方法。基于颜色直方图提取全局颜色特征,利用多尺度空间来表达目标的局部特征,最后将全局和局部特征进行数据融合得到图像的识别结果。实验结果表明,该方法很好地结合了目标的整体和局部信息,能有效地识别目标,且识别效果优于单一的全局特征和局部特征的识别效果。
In this paper,in order to increase the rate of the recognition and speed,we mention a new method to extract and analysis the characteristics of the target.Firstly extraction the global color features from the color histogram.Secondly using multi-scale space to express the local characteristics of target.At last fuse the global and the local feature into the result of the recognition.Experimental results show that this method is good at fuse the global and local information of the target and have a better result of the recognition.
出处
《微型机与应用》
2011年第14期41-43,共3页
Microcomputer & Its Applications
关键词
目标识别
全局特征
局部特征
数据融合
target recognition
global feature
local feature
data fusion
