Based On Digital Image Processing, Flame Combustion Stability Study

Furnace flame diagnosis is a hot issue, but the usual systems have some shortcomings, such as unstable combustion diagnostics, low real-time and recognition rate. Judging the status of the furnace flame accuracy real-time and realizing combustion diagnosis system automate and intelligently can improve the combustion state of recognition accuracy and reliability, and can effectively prevent the flameout, boiler exploding and reduce the loss of state property and protect people's lives.The paper analyses the status of the boiler combustion diagnosis, sums up the research directions at home and abroad, points out their shortcomings, then analyses and summarizes the digital image processing technology and artificial intelligence research, focus on research of combustion diagnostic system based on digital image processing technology and modern artificial intelligence.First of all, we need obtain digital image information about the image acquisition card that includes filtering the original image. Image segmentation is used to diagnose the flame area, and the paper uses the improved two-dimensional histogram image to oblique sub-segmentation algorithm, and then splits out the flame region from which we can extract information reflecting the combustion conditions. After then, we should extract the most representative features through analytical study. Finally, we use intelligent models to diagnose the flame burning state, and test the validity and accuracy of the model. In this paper, we use two intelligent models to determine the stability of the combustion flame image respectively. The neural network model is successfully applied in the boiler combustion system and support vector machine model which is developing rapidly in recent years. We compare their strengths and weaknesses in the paper. A new artificial intelligence discrimination algorithm is proposed for the stability of furnace flame system.At last, the paper points out the shortcomings and the future trend. The advantages of digital imaging technology are no doubt. I believe combustion diagnostics will be more accurate and more sophisticated combining in the digital image processing and artificial intelligence.