基于小波变换的视觉信号特征提取方法  

THE FEATURE EXTRACTION METHOD OF VISUAL SIGNAL BASED ON WAVELET TRANSFORMING

作　　者：张新荣[1] 左建中[2] 张钢[1] 李晓红[1] 

机构地区：[1]天津大学电子信息学院,天津300072 [2]天津大学机械学院,天津300072

出　　处：《电子与信息学报》2003年第4期557-561,共5页Journal of Electronics & Information Technology

摘　　要：该文提出了一种用于视觉信号特征提取的新方法——将CCD获取的图像投影后的信号波形进行小波变换。当小波的尺度选择恰当时,小波变换结果的零交叉点可用于确定图像的中心。实验结果表明,该方法相对于其他方法对随机误差的干扰具有更强的鲁棒性。A new method of feature extraction for visual signal is presented in this paper. The signal of the image projection, which is acquired by CCD, is transformed with wavelet. And the zero cross point of transform is located at the center of the image when the scale of wavelet is suitable. The experiment shows that the method possesses stronger robustness for stochastic error disturbance than other methods do.

关 键 词：小波变换 图像投影 特征提取 零交叉点 视觉信号 芯片 

分 类 号：TN919.8[电子电信—通信与信息系统] O177.6[电子电信—信息与通信工程]