A decision tree algorithm for license plate recognition based on bagging

Decision tree learning is a kind of approximation discrete function value method. It has accurate classification, and is fast-enough for performance. In this paper, a new method of license plate characters recognition is proposed. In this method, the training decision tree classifier based on the bagging theory is put forward on the basis of the license plate characters. Then, the characteristics of license plate character in the image data are extracted. After that, the decision tree classifier is designed. Finally, the extracted feature vector is used in training samples. Experimental results illustrate that the algorithm of license plate recognition is effective and can increase the recognition accuracy distinctly.