License plate Chinese character recognition affected the license plate recognition system is an important technical bottleneck. In this paper, first all sorts of license plate Chinese character recognition algorithms are reviewed, and then license plate Chinese character recognition algorithms based on PCA are studied in depth. For the shortage of the conventional PCA algorithm, license plate Chinese character recognition based on block PCA algorithm is proposed. License plate character images are divided into some smaller non-overlapping sub-images and the PCA algorithm is applied to each of these sub-images. Then the weight sets of local information of Chinese characters are obtained and the features are more representative. For improving the efficiency of PCA algorithm, license plate Chinese character recognition based on block 2DPCA algorithm is proposed. In this method, the partitioned training sub-image samples are as the new training sample images. Both global statistical information and local detailed information of images are considered by the block 2DPCA algorithm. |