摘要
将主成分分析方法(PCA)应用于车牌识别。首先根据采集到样本分类构造各类样本对应特征子空间,然后对待识别字符图片进行预处理,再分别向各类特征空间投影,根据重构误差判断类别识别字符。
The paper proposes principal component analysis (PCA) method application for license plate recognition. PCA constructs corresponding feature subspaces for the sample images and preprocesses the sample character images, then projects the unidentified character image to the feature space and recognises character using classification judgement based on the reconstruct error. Experiments confirm that the PCA method can keep much image information and is robust for character recognition distortion. As a result improved recognition performance is achieved.
出处
《计算机与数字工程》
2007年第3期110-112,共3页
Computer & Digital Engineering
关键词
车牌识别
特征空间
奇异值分解(SVD)
主成分分析(PCA)
License Plate Recognition,feature space,singular value decomposition (SVD),principal component analysis (PCA)
