摘要
提出了一个基于小波变换和CIP算子的数字图像定位和识别系统，该系统用于移动机器人的门牌识别, 首先通过小波变换压缩图像，然后直接对压缩的彩色图像运用CIP算子处理并进行定位，最后用BP神经网络识别。小波变换压缩了图像，从而减小了图像处理的数据量，加快了整个检测定位和识别的速度。CIP算子针对彩色图像的特点，全面作用在Y.I.Q色彩空间的3个分量上，在彩色门牌定位的边缘检测中有很好的效果。实验结果表明了该系统所采用方法的有效性。
In this paper, one digital image detection and recognition system,which is used for the mobile robots recognition for door license, is proposed. After being compressed by wavelet transform, the image of door license is then processed by the CIP operator and recognized by the BP neural networks. The wavelet decreases the data size of the image, thus increases the speed of the total processing greatly. The CIP operator segments the license from the background and positions it accurately, operates on the three channels of the Y.I.Q color space. The experiments results prove the methods in this system are effective.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第13期107-108,194,共3页
Computer Engineering
