Research On The Character Detection And Identification

With the rapid development and applications of the computer and multimedia, digital characters detection and identification have broad application in many areas in today's society. Although many methods have been proposed , many problems still exist when the specific problems are related.The work according to features of the digital characters in this paper is as follows:First, this paper researches Independent Component Analysis(ICA) algorithm of pattern recognition technology and particularly treats basic principle and implementation of ICA, roundly introduces current several main ICA algorithms and the relation of them.analysis a fast algorithm, FastICA. On this basis, this paper research character recognition application using an improved FastICA algorithm. It accelerates the convergence of speed, digital character recognition speed than general algorithm.Second, Two pre-processing technologies for vectorizing scanned topographic maps are presented. Firstly, an algorithm for extraction of connected components based on mathematical morphology is employed to find out elevation value notes; This paper also includes maps pretreatment, such as binary. Binary effects the following work, the effective binary algorithm guarantees a high detection rate. After extracting connected component, dilate candidate number, and structural elements chooses rotundity. It not only eliminate non-digit character elements but also calculate angle of rotation. Secondly, recognize digital character based on the numerical expression of the background.Compar- ed the method of based on the independent component analysis, this method is more simple. Experimental results show that the method has a good experimental effect of various fonts, different font sizes character recognition.