Earticle

초록

영어

Due to the low recognition accuracy, the remote meter reading technology based on camera direct reading has been developed slowly. Although there is a variety of features data for recognizing digit in image using BP neural network, some of data cannot be used to recognize digit accurately. Moreover, the BP network has a slow rate of convergence, low accuracy and easily fall into local minimum. To solve the above questions, a new digit recognition algorithm of meter based on rough set and neural network which are optimized by genetic algorithm is proposed. The improved genetic rough set algorithm is used for reducing the data, and then the minimum feature attribute sets after reduction are input to genetic neural network for identifying digit. The experimental results show that the algorithm can effectively reduce the number of decision attributes and simplify the structure of the neural network with high identification accuracy and short training time, which improve the generalization ability and robustness of the neural network.

목차

Abstract
 1. Introduction
 2. The Basic Concepts of Rough Set Theory
 3. Rough Set Attribute Reduction Algorithm based on Genetic Algorithm Optimization
  3.1. The Selection of Fitness Function
  3.2. The Implementation Process of the Genetic Reduction Algorithm
 4. BP Neural Network based on Genetic Algorithm Optimization
  4.1. The Basic Principles of Genetic Neural Network
  4.2. The Learning Process based on Genetic Algorithm to Optimize the Weights and Thresholds of Neural Network
 5. The Simulation Experiment and Result Analysis of Digit Recognition for Meter Image
  5.1. Image preprocessing
  5.2. Character Feature Extraction
  5.3. Attribute Reduction of the Feature Vector based on Genetic Rough Set
  5.4. Digit Recognition based on the Genetic Neural Network
 6. Conclusion
 Acknowledgements
 References

키워드

저자

• Xiaochen Zhang [ College of Communication Engineering, Chongqing University, Chongqing, 400044, China ]
• Yuanchang Zhong [ College of Communication Engineering, Chongqing University, Chongqing, 400044, China, School of Automation, Chongqing University, Chongqing, 400044, China ]
• Jiajia Shen [ College of Communication Engineering, Chongqing University, Chongqing, 400044, China ]
• Kun Li [ College of Communication Engineering, Chongqing University, Chongqing, 400044, China ]
• Congjun Feng [ College of Communication Engineering, Chongqing University, Chongqing, 400044, China ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Smart Home

• 간기

  격월간

• pISSN

  1975-4094

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605