Research On The Implementation Of The Embedded System For Recognizing Chinese Character Handwriting Based On Qt/Embedded Framework

Handwritten Chinese character recognition is a typical problem in the field of pattern recognition and text recognition.With the rapid development of "Internet of Things" technology,handwritten Chinese character recognition technology has been widely used in various embedded devices.Faced with a wide variety of embedded systems,Qt/Embedded framework is suitable for the development of embedded handwritten Chinese character recognition systems with its good cross-platform features.The basic process of handwritten Chinese character recognition includes preprocessing,feature extraction,classification,etc.For the special environment of embedded system,the corresponding aspects of recognition are improved accordingly.Statistical feature recognition method and high precision classifier are used,and mixed features are adopted.The two-level classification recognition scheme designs and implements a handwritten Chinese character recognition engine.The linearization normalization technique is used to unify the handwritten Chinese character image size for feature extraction and classification.In the first classifier,combined with fuzzy mesh partitioning,a feature extraction method based on pixel density is proposed,and the obtained low-dimensional feature vector is sent.The minimum distance classifier uses the kd tree to calculate 300 candidate Chinese characters.The two-level classifier combined with the elastic mesh division adopts the gradient feature extraction method,and the obtained high-dimensional feature vector is reduced by the LDA algorithm.The reduced-dimensional thick feature vector is sent to the high-precision MQDF classifier,which is among the 300 candidate Chinese characters.Perform classification and identification to obtain the final Chinese character recognition result.Qt/Embedded is a version developed by Qt framework specifically for the embedded environment.By combining Qt/Embedded input method framework and handwritten Chinese character recognition engine,an embedded handwritten Chinese character recognition system using cloud computing is designed and implemented.After testing,the system has high recognition accuracy and fast recognition speed.And the system has a certain application value of embedded environment.