Journal of Applied Sciences ›› 1998, Vol. 16 ›› Issue (1): 81-87.

• Articles • Previous Articles     Next Articles

Study on Neural Network Diagnosis System Based on Characteristics of Flame Image

YANG HONGMIN, TANG XIAOJUN, LIU YONG, GU FAN, XU YIQIAN   

  1. The Thermal Energy Engineering Institute, Southeast University, Nanjing 210096
  • Received:1996-05-16 Revised:1996-09-19 Online:1998-03-31 Published:1998-03-31

Abstract: The image characteristics of gas flame and pulverized coal flame is studied experimentally and theoretically by using computer digital image processing technology. Based on flame characteristic extraction, the theory of neural network is applied to study the flame recognition. The method of flame combustion recognition is proposed and a practical intelligent flame diagnosis system is developed.

Key words: artificial neural network, flame diagnosis, image processing