Title :
Research of the methods of character recognition for landmark encoding
Author :
Liu, Jihong ; He, Rui ; Yang, Li
Author_Institution :
Coll. of Inf. Sci. & Eng., Northeastern Univ., Shenyang, China
Abstract :
The character recognition in landmarks is one of the key technologies in automatic navigation. After introducing the landmark recognition system, the character recognition algorithm based on the multiple perceptron neural network is presented. First of all, the algorithm is simulated in MATLAB, and the weights and thresholds for the sampled landmark character are obtained by training the perceptron neural network. Then, the weights and thresholds are burned to the ROM in the FPGA application board. Finally, in the Quartus II platform, the program based on HDL language is designed to complete the character recognition in this FPGA application board. The experimental results show that the system implements the real-time character recognition with high stability and robustness.
Keywords :
character recognition; computer vision; encoding; field programmable gate arrays; hardware description languages; learning (artificial intelligence); mathematics computing; perceptrons; FPGA application board; HDL language; MATLAB; Quartus II platform; ROM; automatic navigation; landmark encoding; landmark recognition system; multiple a network; perceptron neural network; real-time character recognition; sampled landmark character; training; Character recognition; Clocks; Educational institutions; Field programmable gate arrays; Hardware; Random access memory; Read only memory; Landmark character recognition; monitoring pattern classifier; perception network;
Conference_Titel :
Industrial Electronics and Applications (ICIEA), 2012 7th IEEE Conference on
Conference_Location :
Singapore
Print_ISBN :
978-1-4577-2118-2
DOI :
10.1109/ICIEA.2012.6361055